feat(web): mode-aware backlog banner — stop telling manual users summaries auto-arrive

The first pilot user sat in Manual mode reading "new summaries land gradually,
check back tomorrow" — copy that only makes sense in Automatic mode. Manual mode
never auto-summarizes, so the banner promised delivery that would never come.

The list page now reads the user's summarize mode and shows mode-correct copy:
- Auto: unchanged "land gradually" backlog note.
- Manual: "new videos appear here but are not summarized automatically — use the
  Summarize button" plus a "Switch to Automatic" link to /account.
The connected-but-empty first-run state is likewise mode-aware.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

CLAUDE.md

@@ -46,8 +46,9 @@ These caused real mistakes that were caught and corrected; the corrections are l
 (See `DECISIONS.md` for full rationale. Listed here so you don't propose them.)
 
 - **No Supabase** — reuse Dex / ESO+1Password / Postgres (ADR-002).
-- **No global cross-tenant video/transcript table** — per-user isolation (data-model). Dedup
+- **No global cross-tenant *video* table** — videos stay per-user (data-model). Transcripts ARE
-  across users is a Future C concern, not a Stage 0/1 default.
+  shared since ADR-021 (public caption content, keyed by `(provider, provider_video_id)`, non-RLS)
+  so re-analysis never re-fetches; the *videos* half of cross-tenant dedup stays a Future C concern.
 - **No audio-download + speech-to-text in the core path** — captions-first (ADR-007). STT is a
   deferred, bounded optional component.
 - **No public SaaS / sign-up / billing / Google OAuth verification at scale** — Future C,
@@ -60,8 +61,14 @@ These caused real mistakes that were caught and corrected; the corrections are l
   stores, and sinks are adapters. Adding a video provider or a sink = a new adapter implementing
   the interface, nothing in the engine changes. This is what keeps "standalone vs homelab" a
   wiring choice (ADR-003).
-- **BDD.** The `docs/use-cases/*.feature` files are the behavior spec. New behavior gets a
+- **BDD.** The `docs/use-cases/*.feature` files are the behavior spec (design records — there is
-  scenario; the use-case core is tested through fake adapters, not live YouTube/brain.
+  no godog runner). New behavior gets a scenario; the use-case core is tested through fake
+  adapters, not live YouTube/brain. A name-coverage gate (`test/acceptance/scenario_coverage_test.go`,
+  `TestScenarioCoverage`) keeps the two from drifting: every non-`@pending` scenario must be
+  mapped to an existing Go test in `scenarioCoverage`. When you add a scenario, either map it to
+  its covering test or tag it `@pending` in the `.feature` with a one-line reason. It checks the
+  *link*, not that the test exercises the scenario — that's the deliberate trade for not running
+  godog (see issue #5 / the BDD-runner decision).
 
 ## Skills (engineering discipline)
 
@@ -79,7 +86,7 @@ Skills live in the canonical library `mathias/skills` and are wired into this re
 
 ## Current build state (start here for the first task)
 
-The repo is **green and shipping** — last tag `v0.4.0`. `task check` passes (fmt, vet, lint,
+The repo is **green and shipping** — last tag `v0.15.0`. `task check` passes (fmt, vet, lint,
 `go test -p 1 ./...`). Go is `1.26.1` (see `go.mod`).
 
 - Clean Architecture core is implemented: `internal/domain` (entities), `internal/ports`
@@ -88,13 +95,19 @@ The repo is **green and shipping** — last tag `v0.4.0`. `task check` passes (f
   green against it.
 - Adapters present under `internal/adapters/`: `youtube` (captions-first `VideoSource`,
   timedtext/InnerTube acquisition per ADR-010), `summarizer` + `llm` (the copied AI router,
-  Primary→Fallback per ADR-004), `store` (Postgres, golang-migrate migrations 001–006),
+  now a resilient endpoint chain — local primary → local fallback → external worst-case,
+  parse-failure-aware, ADR-004 + ADR-022), `store` (Postgres, golang-migrate migrations 001–015),
   `secrets` (file-backed `SecretStore`). The brain HTTP sink (ADR-005) is the remaining
   optional sink.
 - Stage 1 is open (ADR-012): multi-user with **DB-enforced** isolation — Postgres RLS `FORCE`d
   on all user-owned tables (migration 003), two-user isolation test in
   `internal/adapters/store/rls_test.go`. Registration gate, per-user YouTube web connect, and
   account management (disconnect / delete, ADR-013) all shipped.
+- Transcript persistence (ADR-021, migration 015): transcripts are a **shared, non-RLS** store
+  keyed by `(provider, provider_video_id)` — the single exception to the isolation boundary
+  (`TestTranscriptsTableIsSharedNotRLS`). The engine reads stored transcripts before any caption
+  fetch (`usecase.resolveTranscript`), so re-analysis — re-summarize, paste-a-URL, onboarding
+  burst — never re-touches YouTube. Per-user summaries/videos stay RLS-scoped.
 - `cmd/tapir` subcommands: `list`, `show`, `auth` (interactive host-side OAuth), `run` (batch
   watch→summarize), `serve` (the HTMX+Templ web reader/writer under `internal/web`, a new
   transport over the unchanged engine/ports — ADR-003). `tapir env` prints config.

DECISIONS.md

@@ -743,6 +743,118 @@ collapse keys off the same window (`App.RecencyWindow=0` → everything inline).
 
 ---
 
+## ADR-021 — Persist transcripts as shared, video-keyed public content (re-analysis never re-fetches)
+
+**Status:** Accepted (2026-06-09). **Reopens the transcripts half of** the "Global cross-tenant
+`videos`/`transcripts` table" rejection (data-model.md). **Builds on ADR-007** (captions-first),
+**ADR-010/ADR-014** (the per-IP caption rate gate), and **ADR-012** (per-user RLS isolation).
+
+**Context.** Every summarization fetches the transcript fresh through the caption path, even when
+the exact same transcript was fetched moments ago — for the same user re-summarizing, or for a
+second user who happens to watch the same video. The caption fetch is the one genuinely scarce,
+genuinely risky operation in the system: YouTube's timedtext endpoint is unofficial and per-IP
+rate-limited (ADR-010), and tripping it risks the maintainer's Google standing (ADR-014). So the
+operation we most want to *avoid repeating* is the one we currently repeat unconditionally. A
+transcript is **public content** — the same words YouTube serves to anyone — and carries nothing
+user-identifying. The per-user isolation that protects summaries, feeds, and tokens (ADR-012) is
+the wrong shape for it: it forces a re-fetch per user for data that is identical across users.
+
+The original rejection ("Global cross-tenant `videos`/`transcripts` table") bundled videos and
+transcripts together and rejected both on the grounds that "at 1–5 users, re-summarizing is
+cheaper than the coupling." That reasoning holds for **videos** (per-user feed rows, genuinely
+user-scoped) but not for **transcripts**: the cost being avoided is not LLM re-summarization, it
+is a *rate-gated, reputation-risky network fetch*, and that cost is paid per re-fetch regardless
+of user count. One re-fetch avoided is strictly worth more than the coupling it removes.
+
+**Decision.**
+1. **A single shared `transcripts` table, keyed by the cross-user dedup key
+   `(provider, provider_video_id)`** — the stable public identity of the video, not Tapir's
+   internal per-user `videos.id`. Columns: the key, `source` (`captions`/`none`), `language`,
+   `content`, `fetched_at`. It holds **only public caption content + the video's public id** —
+   nothing user-identifying — and is therefore **NOT RLS-scoped**: no `user_id`, no policy, no
+   `FORCE ROW LEVEL SECURITY`. This is the deliberate, single exception to the ADR-012 isolation
+   boundary, and the only one.
+2. **Summarize path becomes read-stored-first.** Have a stored transcript for this video? →
+   summarize from the stored text, **no caption fetch**. No stored transcript? → fetch *through
+   the unchanged gate* (ADR-014) → store it → summarize. The gate is neither bypassed nor
+   weakened; persistence reduces how *often* we reach it, never how *fast*.
+3. **De-facto cross-user dedup is the intended behaviour, not a feature with a switch.** Two
+   users who share a video share the one transcript row. A permanent `source = 'none'` (no
+   captions) is stored too, so a known-caption-less video is not re-fetched by anyone. A
+   transient 429 (`SourceRateLimited`) is **never** stored as terminal — it stays a per-user
+   retry via the existing `transcript_status` backoff (ADR-014), so persistence cannot mask a
+   rate-limit into a false "no transcript."
+4. **Per-user `summaries` stay RLS-scoped (ADR-012 unchanged)** and reference the transcript by
+   video id. Videos stay per-user. Only transcripts go shared.
+
+**Consequences.** Re-analysis (re-summarize, different model, paste of an already-seen video,
+onboarding of a second user with overlapping subscriptions) never re-touches YouTube — the
+primary win, and it *reduces* aggregate caption-gate pressure, reinforcing ADR-010/ADR-014 rather
+than straining them. The isolation surface gains exactly one non-RLS table; an isolation test
+asserts the boundary is *exactly* there and has not leaked to any user-owned table (this is the
+proof the public-content classification was implemented as designed). It also unblocks
+multi-model / customizable analysis (re-run analysis on stored text for free) — enabling that is
+this ADR's point; building it is separate.
+
+**Reversibility.** The read-stored-first check is the only behavioural coupling; removing it
+restores fetch-every-time. The down-migration recreates the per-user RLS-scoped transcripts shape
+(001/003). No user-facing surface depends on cross-user sharing — sharing is the *storage shape*,
+never exposed in the UI.
+
+---
+
+## ADR-022 — Summarizer is a resilient endpoint chain, not a single model
+
+**Status:** Accepted (2026-06-10). **Extends ADR-004** (the copied `llm` Primary→Fallback
+routing). Triggered by the first friendly-pilot live run, where a connected user got **zero**
+summaries after 12h.
+
+**Context.** Stage-0 ran a single summarizer model (`koala/phi4-mini`) with no fallback wired
+(`summarizer.New(primary, nil)`). The live run exposed three independent failure modes, each of
+which silently produced no summary:
+
+1. **Context overflow.** `phi4-mini` has an 8k context. Real transcripts (one was 11,602 tokens)
+   exceed it and the gateway returns HTTP 400 — and the request also sent `max_tokens=8192`, so
+   even a short transcript plus the completion budget could overflow the window.
+2. **Malformed model output.** `phi4-mini` intermittently emits `highlights` as a bare string
+   instead of an array, producing `cannot unmarshal string into []string`. The old code returned
+   the parse error **without** trying any other model — a 200-with-bad-JSON short-circuited.
+3. **No fallback existed at all** — any primary failure was terminal for that video.
+
+`phi4-mini` is kept as primary deliberately: it is fast and, on transcripts that fit, correct.
+The fix is resilience around it, not replacing it.
+
+**Decision.**
+1. **Ordered endpoint chain (`summarizer.NewChain`).** Endpoints are tried in order; the first to
+   return a *parseable* summary wins. Default chain:
+   `koala/phi4-mini` (primary, local) → `koala/phi4-14b` (fallback, local) →
+   `berget/mistral-small` (worst-case, external). All three are reached through the **one** LiteLLM
+   gateway by alias — the gateway already fronts both llama-swap and berget — so a fallback is a
+   different alias, not a second client config.
+2. **A parse failure advances the chain, same as a transport error.** "Reliably summarized" means
+   *parseable summary returned*, not *HTTP 200*. This is the behaviour the old Primary→Fallback
+   shape missed.
+3. **Tolerant parse.** `highlights`/`takeaways` coerce from a bare string (or a mixed scalar
+   array) to `[]string`, so the most common small-model quirk is absorbed **without** spending a
+   fallback round-trip — keeping the fast path fast.
+4. **Transcript truncation (`TAPIR_MAX_TRANSCRIPT_CHARS`, default 18000).** Input is bounded
+   up-front to fit a small-context primary, so overflow is prevented rather than recovered-from.
+5. **Bounded completion budget (`TAPIR_SUMMARY_MAX_TOKENS`, default 1500).** A summary needs few
+   hundred tokens; the old 8192 budget itself contributed to 8k-window overflow.
+
+**Local-first guarantee preserved.** The chain ordering *is* the guarantee: locals are tried
+first, so content reaches the external endpoint only after every local endpoint has failed.
+`TAPIR_CLOUD_FALLBACK_MODEL=""` removes the external endpoint entirely — the lever a
+**client/NDA deployment** pulls so content never leaves the local stack. With no external endpoint
+configured the `ai_routing.feature` "content only local" scenarios hold unchanged.
+
+**Reversibility.** Pure wiring + config. Setting `TAPIR_FALLBACK_MODEL` and
+`TAPIR_CLOUD_FALLBACK_MODEL` empty collapses the chain back to single-primary behaviour; the
+tolerant parse and truncation are strict supersets of the old behaviour (a previously-parseable
+reply still parses; a transcript within budget is unchanged).
+
+---
+
 ## Rejected alternatives
 
 Approaches considered during the 2026-06-02 planning + grill session and **deliberately not
@@ -757,7 +869,7 @@ maps to the ADR that settles it.
 | Lifting shared packages into a `brain-common` module | Couples Tapir's release cycle to the monolith for negligible code savings | ADR-004 |
 | Importing/replicating the filesystem `brain` package | Assumes co-location with the brain git checkout; wrong for a standalone networked service | ADR-005 |
 | Reusing `ingestion`'s `oauth` package for YouTube/Vimeo | Same name, opposite direction — it's inbound MCP-server auth, not outbound provider OAuth | ADR-006 |
-| Global cross-tenant `videos`/`transcripts` table (dedup) | Reintroduces the cross-domain DB coupling the homelab review is removing; at 1–5 users, re-summarizing is cheaper than the coupling | data-model.md |
+| Global cross-tenant `videos`/`transcripts` table (dedup) | Reintroduces the cross-domain DB coupling the homelab review is removing; at 1–5 users, re-summarizing is cheaper than the coupling. **Transcripts half reopened by ADR-021** — the avoided cost there is a rate-gated, reputation-risky *caption fetch*, not LLM re-summarization, so it outweighs the coupling; **videos stay per-user.** | data-model.md, **ADR-021** (transcripts only) |
 | Audio-download + Whisper STT in the core path | ToS-grey, breakage-prone (yt-dlp), contends for koala GPU with the JEPA PoC; captions alone test the core hypothesis | ADR-007 |
 | Building multi-tenant SaaS / Google OAuth verification now | "Real users soon" was lowered to Future B; SaaS machinery before the Stage 0 self-use gate is the primary documented anti-goal | ADR-008, VISION |
 | Delegating the S5 reuse spike to an agent swarm | A 1-hour sequential read-and-judge with a single coupled conclusion; orchestration overhead exceeds the work, and it's Diamond-1 judgment the maintainer wanted to own | (process note) |

README.md

@@ -75,8 +75,10 @@ password or Google); a new Dex subject is routed to `/register` to create a Tapi
 Stage 1 is multi-user: each user connects their own YouTube account from the browser and
 manages their own summaries under DB-enforced RLS isolation. When `TAPIR_DISCOVERY_INTERVAL`
 is set (e.g. `2h`), the serve process runs a scheduled discovery pass for every registered
-user automatically — no CronJob required. See `docs/homelab-integration.md` for the full
+user automatically — no CronJob required. In auto mode only videos published within
-config reference.
+`TAPIR_AUTO_SUMMARIZE_WINDOW` (default ~7d, ADR-020) are summarised automatically; older videos
+are listed and summarised on demand, so a large back-catalogue doesn't keep re-driving the
+caption rate gate. See `docs/homelab-integration.md` for the full config reference.
 
 ### Headless on koala
 

cmd/tapir/discovery.go

@@ -0,0 +1,50 @@
+package main
+
+import (
+	"context"
+	"log/slog"
+	"sync"
+
+	"gitea.d-ma.be/mathias/tapir/internal/runner"
+)
+
+// discoveryRunner runs one user's discovery pass.
+type discoveryRunner func(ctx context.Context, userID string) (runner.Stats, error)
+
+// serialize wraps run so calls never overlap: every discovery pass — scheduled
+// or connect-triggered (#6) — acquires the same lock, preserving the
+// one-fetcher-at-a-time invariant the scheduler relies on (ADR-018, the
+// single-replica assumption). Locking is per-user, so a connect-triggered pass
+// interleaves between the scheduler's users instead of waiting for a whole pass.
+func serialize(mu *sync.Mutex, run discoveryRunner) discoveryRunner {
+	return func(ctx context.Context, userID string) (runner.Stats, error) {
+		mu.Lock()
+		defer mu.Unlock()
+		return run(ctx, userID)
+	}
+}
+
+// discoveryTrigger fires an out-of-band discovery pass for one user without
+// blocking the caller (the connect HTTP handler). The pass runs on the server's
+// long-lived ctx — not the request ctx — so it survives the post-connect
+// redirect. run is the serialized runner, so a trigger never overlaps the
+// scheduler. Satisfies web.DiscoveryTrigger.
+type discoveryTrigger struct {
+	ctx context.Context
+	run discoveryRunner
+	// onboard, when set, runs after the discovery pass to summarize a capped number
+	// of the user's newest videos (Feature 1). Optional.
+	onboard func(ctx context.Context, userID string)
+	log     *slog.Logger
+}
+
+func (t *discoveryTrigger) Enqueue(userID string) {
+	go func() {
+		if _, err := t.run(t.ctx, userID); err != nil {
+			t.log.Warn("discovery: connect-triggered pass had errors", "user", userID, "err", err)
+		}
+		if t.onboard != nil {
+			t.onboard(t.ctx, userID)
+		}
+	}()
+}

cmd/tapir/discovery_test.go

@@ -0,0 +1,62 @@
+package main
+
+import (
+	"context"
+	"fmt"
+	"sync"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+
+	"gitea.d-ma.be/mathias/tapir/internal/runner"
+)
+
+// serialize must guarantee at most one discovery pass runs at a time, so a
+// connect-triggered pass never fetches concurrently with the scheduler.
+func TestSerializeRunsOneAtATime(t *testing.T) {
+	var active, maxActive int32
+	run := func(_ context.Context, _ string) (runner.Stats, error) {
+		n := atomic.AddInt32(&active, 1)
+		for { // record the high-water mark of concurrent runs
+			m := atomic.LoadInt32(&maxActive)
+			if n <= m || atomic.CompareAndSwapInt32(&maxActive, m, n) {
+				break
+			}
+		}
+		time.Sleep(2 * time.Millisecond)
+		atomic.AddInt32(&active, -1)
+		return runner.Stats{}, nil
+	}
+
+	s := serialize(&sync.Mutex{}, run)
+	var wg sync.WaitGroup
+	for i := 0; i < 20; i++ {
+		wg.Add(1)
+		go func(i int) { defer wg.Done(); _, _ = s(context.Background(), fmt.Sprintf("u%d", i)) }(i)
+	}
+	wg.Wait()
+
+	require.Equal(t, int32(1), atomic.LoadInt32(&maxActive),
+		"serialize must run at most one pass at a time")
+}
+
+// Enqueue runs the user's pass out-of-band (non-blocking) on the trigger's ctx.
+func TestDiscoveryTriggerEnqueueRunsUser(t *testing.T) {
+	done := make(chan string, 1)
+	run := func(_ context.Context, userID string) (runner.Stats, error) {
+		done <- userID
+		return runner.Stats{}, nil
+	}
+	tr := &discoveryTrigger{ctx: context.Background(), run: run, log: quietLog()}
+
+	tr.Enqueue("u1")
+
+	select {
+	case got := <-done:
+		require.Equal(t, "u1", got)
+	case <-time.After(2 * time.Second):
+		t.Fatal("Enqueue did not run the user's pass")
+	}
+}

cmd/tapir/main.go

@@ -20,6 +20,7 @@ import (
 	"net/http"
 	"os"
 	"os/signal"
+	"sync"
 	"time"
 
 	"gitea.d-ma.be/mathias/tapir/internal/adapters/secrets"
@@ -209,7 +210,10 @@ func cmdServe(ctx context.Context, log *slog.Logger) error {
 			ClientSecret: cfg.YTClientSecret,
 			RedirectURL:  cfg.YTConnectRedirectURL,
 		}, secretStore, st, log)
-		log.Info("web youtube connect enabled", "redirect", cfg.YTConnectRedirectURL)
+		// Paste-a-URL (Feature 2): same YouTube credentials, per-user adapter built
+		// per request. Mounting the /paste route keys off app.Fetcher being set.
+		app.Fetcher = videoFetcher{cfg: cfg, secrets: secretStore}
+		log.Info("web youtube connect + paste enabled", "redirect", cfg.YTConnectRedirectURL)
 	}
 
 	// Immediate summarization for the web "Summarize" button. When the engine can
@@ -238,13 +242,44 @@ func cmdServe(ctx context.Context, log *slog.Logger) error {
 	if cfg.DiscoveryInterval > 0 {
 		log.Info("scheduled discovery enabled", "interval", cfg.DiscoveryInterval, "fetch_rate", cfg.FetchRate)
 		log.Warn("scheduled discovery assumes a SINGLE replica — running serve at >1 replica double-runs discovery (ADR-018)")
-		runUser := func(ctx context.Context, userID string) (runner.Stats, error) {
+		rawRunUser := func(ctx context.Context, userID string) (runner.Stats, error) {
 			r, err := buildUserRunner(cfg, st, secretStore, userID, log)
 			if err != nil {
 				return runner.Stats{}, err
 			}
 			return r.RunOnce(ctx)
 		}
+		// One lock shared by the scheduler and connect-triggered passes (#6) so
+		// they never fetch concurrently — the single-fetcher invariant (ADR-018).
+		runUser := serialize(&sync.Mutex{}, rawRunUser)
+		// Onboarding burst (Feature 1): after the connect-triggered discovery pass,
+		// summarize up to OnboardSummarizeCount of the user's NEWEST unsummarized
+		// videos so a fresh account gets real summaries in its first session. Hard
+		// cap; explicit, so it bypasses the recency window — but every fetch still
+		// goes through globalFetchGate via the Processor. No-op when disabled
+		// (count 0) or queue-only (no Processor).
+		onboard := func(ctx context.Context, userID string) {
+			if cfg.OnboardSummarizeCount <= 0 || app.Processor == nil {
+				return
+			}
+			ids, err := st.NewestUnsummarizedVideoIDs(ctx, userID, cfg.OnboardSummarizeCount)
+			if err != nil {
+				log.Warn("onboarding: list newest unsummarized", "user", userID, "err", err)
+				return
+			}
+			for _, id := range ids {
+				if err := app.Processor.ProcessVideo(ctx, userID, id); err != nil {
+					log.Warn("onboarding: summarize", "user", userID, "video", id, "err", err)
+				}
+			}
+			if len(ids) > 0 {
+				log.Info("onboarding burst complete", "user", userID, "summarized", len(ids), "cap", cfg.OnboardSummarizeCount)
+			}
+		}
+		if app.Connect != nil {
+			app.Connect.Discovery = &discoveryTrigger{ctx: ctx, run: runUser, onboard: onboard, log: log}
+			log.Info("connect-triggered discovery enabled", "onboard_cap", cfg.OnboardSummarizeCount)
+		}
 		go runScheduler(ctx, cfg.DiscoveryInterval, st, runUser, log)
 	} else {
 		log.Info("scheduled discovery disabled (TAPIR_DISCOVERY_INTERVAL unset or 0)")

cmd/tapir/processor.go

@@ -3,6 +3,7 @@ package main
 import (
 	"context"
 	"fmt"
+	"strings"
 
 	"gitea.d-ma.be/mathias/tapir/internal/adapters/llm"
 	"gitea.d-ma.be/mathias/tapir/internal/adapters/secrets"
@@ -11,9 +12,29 @@ import (
 	"gitea.d-ma.be/mathias/tapir/internal/adapters/youtube"
 	"gitea.d-ma.be/mathias/tapir/internal/config"
 	"gitea.d-ma.be/mathias/tapir/internal/domain"
+	"gitea.d-ma.be/mathias/tapir/internal/ports"
 	"gitea.d-ma.be/mathias/tapir/internal/usecase"
+	"gitea.d-ma.be/mathias/tapir/internal/web"
 )
 
+// videoFetcher adapts the YouTube adapter to web.VideoFetcher for the paste flow
+// (Feature 2). It builds a per-user adapter bound to that user's token ref and
+// resolves a single video's metadata via the Data API — ungated; only the later
+// transcript fetch goes through globalFetchGate.
+type videoFetcher struct {
+	cfg     config.Config
+	secrets ports.SecretStore
+}
+
+func (f videoFetcher) FetchVideo(ctx context.Context, userID, videoID string) (domain.Video, error) {
+	a := youtube.New(youtube.Config{
+		ClientID:       f.cfg.YTClientID,
+		ClientSecret:   f.cfg.YTClientSecret,
+		TokenSecretRef: web.YouTubeTokenRef(userID),
+	}, f.secrets)
+	return a.VideoByID(ctx, userID, videoID)
+}
+
 // buildProcessor wires the summarization engine — YouTube source (captions-first),
 // AI-router summarizer, store sink — shared by `tapir run` and the web
 // "Summarize now" path so the wiring lives in one place. It returns (nil, nil) —
@@ -22,6 +43,40 @@ import (
 // queue-only fallback: the web UI keeps working (the button just queues) and
 // `tapir run` reports the gap via its own ValidateForRun. Missing engine config
 // is never an error here.
+// buildSummarizer wires the summarization endpoint chain (ADR-022) shared by the
+// web "Summarize now" path and the scheduler's per-user runners. The chain is:
+// primary (local, fast) → local fallback → cloud fallback (worst case). Each
+// endpoint reaches the same LiteLLM gateway with a different model alias — the
+// gateway fronts both llama-swap and berget — so a fallback is just a different
+// alias, not a second client config. Empty model entries are skipped, so a
+// client deployment can set the cloud fallback empty to keep content local.
+func buildSummarizer(cfg config.Config) *summarizer.Summarizer {
+	mk := func(model string) summarizer.Endpoint {
+		return summarizer.Endpoint{
+			Client:   llm.New(cfg.GatewayURL, cfg.GatewayKey, model, cfg.SummarizerTimeout, llm.WithMaxTokens(cfg.SummaryMaxTokens)),
+			Provider: providerOf(model),
+			Model:    model,
+		}
+	}
+	eps := []summarizer.Endpoint{mk(cfg.SummarizerModel)}
+	if cfg.FallbackModel != "" && cfg.FallbackModel != cfg.SummarizerModel {
+		eps = append(eps, mk(cfg.FallbackModel))
+	}
+	if cfg.CloudFallbackModel != "" && cfg.CloudFallbackModel != cfg.SummarizerModel {
+		eps = append(eps, mk(cfg.CloudFallbackModel))
+	}
+	return summarizer.NewChain(eps, cfg.MaxTranscriptChars)
+}
+
+// providerOf maps a model alias to the domain AIProvider recorded on summaries.
+// A "berget/" alias is an external provider; everything else is the local stack.
+func providerOf(model string) string {
+	if strings.HasPrefix(model, "berget/") {
+		return "berget"
+	}
+	return "local"
+}
+
 func buildProcessor(cfg config.Config, st *store.Store) (*usecase.Engine, error) {
 	if cfg.GatewayURL == "" || cfg.YTClientID == "" || cfg.YTClientSecret == "" || cfg.SecretsFile == "" {
 		return nil, nil
@@ -35,15 +90,14 @@ func buildProcessor(cfg config.Config, st *store.Store) (*usecase.Engine, error)
 		PreferredLanguages: []string{"en"},
 	}, secretStore)
 
-	// Local Primary only; no BYO fallback for the demo (fallback nil).
+	sum := buildSummarizer(cfg)
-	primary := summarizer.Endpoint{
-		Client:   llm.New(cfg.GatewayURL, cfg.GatewayKey, cfg.SummarizerModel, cfg.SummarizerTimeout),
-		Provider: "local",
-		Model:    cfg.SummarizerModel,
-	}
-	sum := summarizer.New(primary, nil)
 
-	return usecase.NewEngine(src, sum, st), nil
+	// The store is both the summary sink and the shared transcript cache (ADR-021):
+	// the engine reads stored transcripts before any caption fetch and writes
+	// resolved ones back, so re-analysis never re-touches YouTube.
+	eng := usecase.NewEngine(src, sum, st)
+	eng.Transcripts = st
+	return eng, nil
 }
 
 // engineProcessor adapts the engine (which works in terms of a domain.Video) to

cmd/tapir/scheduler.go

@@ -6,9 +6,7 @@ import (
 	"log/slog"
 	"time"
 
-	"gitea.d-ma.be/mathias/tapir/internal/adapters/llm"
 	"gitea.d-ma.be/mathias/tapir/internal/adapters/store"
-	"gitea.d-ma.be/mathias/tapir/internal/adapters/summarizer"
 	"gitea.d-ma.be/mathias/tapir/internal/adapters/youtube"
 	"gitea.d-ma.be/mathias/tapir/internal/config"
 	"gitea.d-ma.be/mathias/tapir/internal/ports"
@@ -37,22 +35,19 @@ func buildUserRunner(cfg config.Config, st *store.Store, secretStore ports.Secre
 		PreferredLanguages: []string{"en"},
 	}, secretStore)
 
-	primary := summarizer.Endpoint{
+	engine := usecase.NewEngine(src, buildSummarizer(cfg), st)
-		Client:   llm.New(cfg.GatewayURL, cfg.GatewayKey, cfg.SummarizerModel, cfg.SummarizerTimeout),
-		Provider: "local",
-		Model:    cfg.SummarizerModel,
-	}
-	engine := usecase.NewEngine(src, summarizer.New(primary, nil), st)
 
 	return runner.New(src, st, engine, userID, log,
 		runner.WithBackoff(cfg.FetchBackoff),
 		runner.WithAutoWindow(cfg.AutoSummarizeWindow)), nil
 }
 
-// userLister enumerates every registered user. *store.Store satisfies it via
+// userLister enumerates every registered user and reports a user's video
-// ListAllUsers. A small local interface keeps the scheduler testable with a fake.
+// connections. *store.Store satisfies it via ListAllUsers + ConnectionsForUser.
+// A small local interface keeps the scheduler testable with a fake.
 type userLister interface {
 	ListAllUsers(ctx context.Context) ([]store.UserIdentity, error)
+	ConnectionsForUser(ctx context.Context, userID string) ([]store.Connection, error)
 }
 
 // runDiscoveryPass runs one discovery pass for every user. runUser performs a
@@ -78,6 +73,18 @@ func runDiscoveryPass(
 		if ctx.Err() != nil {
 			break // shutting down: stop enumerating
 		}
+		// Skip users with no video connection. A discovery pass for them only
+		// attempts to resolve a token that was never minted, logging a spurious
+		// "ref not found" every tick (e.g. stale Dex-era orphan identities).
+		conns, err := lister.ConnectionsForUser(ctx, u.UserID)
+		if err != nil {
+			log.Warn("scheduler: list connections failed", "user", u.UserID, "err", err)
+			continue
+		}
+		if len(conns) == 0 {
+			log.Debug("scheduler: skipping user with no video connections", "user", u.UserID)
+			continue
+		}
 		stats, err := runUser(ctx, u.UserID)
 		total = sumStats(total, stats)
 		if err != nil {

cmd/tapir/scheduler_test.go

@@ -23,12 +23,23 @@ func quietLog() *slog.Logger {
 type fakeLister struct {
 	users  []store.UserIdentity
 	err    error
+	noConn map[string]bool // users that have NOT connected a video source
 }
 
 func (f fakeLister) ListAllUsers(context.Context) ([]store.UserIdentity, error) {
 	return f.users, f.err
 }
 
+// ConnectionsForUser reports a single youtube connection for every user except
+// those in noConn, which return zero — the connection-less case the scheduler
+// must skip instead of running (and failing to resolve a token for).
+func (f fakeLister) ConnectionsForUser(_ context.Context, userID string) ([]store.Connection, error) {
+	if f.noConn[userID] {
+		return nil, nil
+	}
+	return []store.Connection{{Provider: "youtube"}}, nil
+}
+
 // countingRunUser records how many passes each user got, optionally failing for
 // specific users, under a mutex so it is safe across the scheduler goroutine.
 type countingRunUser struct {
@@ -91,6 +102,21 @@ func TestDiscoveryPassRunsEveryUserOnce(t *testing.T) {
 	require.Equal(t, 3, stats.Summarized, "stats are summed across users")
 }
 
+func TestDiscoveryPassSkipsUsersWithoutConnections(t *testing.T) {
+	// b never connected a video source (e.g. a stale Dex-era orphan identity).
+	// It must be skipped silently — not run and logged as a token error every pass.
+	lister := fakeLister{users: usersN("a", "b", "c"), noConn: map[string]bool{"b": true}}
+	rc := newCountingRunUser()
+
+	stats := runDiscoveryPass(context.Background(), lister, rc.run, quietLog())
+
+	require.Equal(t, 1, rc.count("a"))
+	require.Equal(t, 0, rc.count("b"), "a user with no connection must be skipped, not run")
+	require.Equal(t, 1, rc.count("c"))
+	require.Equal(t, 2, stats.Summarized, "only connected users contribute")
+	require.Equal(t, 0, stats.Errors, "skipping is silent — no spurious error stat")
+}
+
 func TestDiscoveryPassOneUserFailureDoesNotStopOthers(t *testing.T) {
 	lister := fakeLister{users: usersN("a", "b", "c")}
 	rc := newCountingRunUser("b") // user b's pass errors

docs/architecture/architecture.md

@@ -147,9 +147,17 @@ graph TB
   rate-limiting — ADR-014).
 - **Summarization mode** — `users.auto_summarize` (migration 006). Default is **true** for new
   users (migration 011, ADR-018); all existing rows were back-filled via migration 012. Auto:
-  every new video is summarized. Manual: new videos appear unsummarized; the button sets
+  new videos **published within the recency window** (`TAPIR_AUTO_SUMMARIZE_WINDOW`, default ~7d,
-  `videos.summarize_requested`, which the next `tapir run` processes and clears. Both the click
+  ADR-020) are summarized automatically; older videos are discovered and listed but wait for an
-  path and the batch `tapir run` drive the same unchanged engine.
+  explicit "Summarize". Manual: new videos appear unsummarized; the button sets
+  `videos.summarize_requested`, which the next `tapir run` processes and clears. A manual request
+  bypasses the recency bound. Both the click path and the batch `tapir run` drive the same
+  unchanged engine.
+- **List surface (ADR-020)** — the list reads `ListVideos` ordered summarized-first, then
+  `published_at DESC NULLS LAST`. The web layer collapses the noise so summaries are not buried:
+  un-summarized videos older than the recency window fold into one "Show N older videos"
+  disclosure, and caption-less videos collapse to a single count line. Copy surfaces scarcity
+  honestly (queue counts, gradual-fill note) — it never implies the feed is fuller than it is.
 
 The engine, ports, and sink adapters are **untouched** by all of the above — the web surface only
 reads the store and triggers the existing engine. Adding it changed wiring, not the core (ADR-003).
@@ -174,6 +182,7 @@ sequenceDiagram
         loop per user
             S->>DB: GetAutoSummarize(userID)
             S->>YT: ListSubscriptions + NewVideos
+            Note over S,DB: auto: skip videos published before<br/>TAPIR_AUTO_SUMMARIZE_WINDOW (ADR-020);<br/>older ones listed, await manual request
             Note over S,YT: WaitFetchGate(ctx) throttles<br/>all fetches to TAPIR_FETCH_RATE
             alt transcript available
                 S->>LLM: Summarize
@@ -214,19 +223,22 @@ Both paths share `globalFetchGate` — rate limiting is **respected in both**, n
 
 | Path | Trigger | Order | Rationale |
 |------|---------|-------|-----------|
-| **Foreground** | User clicks "Summarize now" on any non-summarized card (`POST /v/{id}/retry-now` for rate-limited; `POST /v/{id}/summarize` for pending) | Single chosen video | On-demand value: user picks a specific video to read now |
+| **Foreground** | User clicks "Summarize" on any non-summarized card (`POST /v/{id}/retry-now` for rate-limited; `POST /v/{id}/summarize` for pending) | Single chosen video | On-demand value: user picks a specific video to read now — bypasses the recency bound |
-| **Background batch** | Scheduled discovery pass every `TAPIR_DISCOVERY_INTERVAL` | **Newest-first across all channels** (see below) | Onboarding prioritisation: most recent, relevant videos surface first |
+| **Background batch** | Scheduled discovery pass every `TAPIR_DISCOVERY_INTERVAL` | **Newest-first across all channels** (see below), **bounded to the recency window** (ADR-020) | Onboarding prioritisation within bounded load: recent videos auto-fill; the older back-catalogue stays on-demand |
 
-The rationale for both paths is **onboarding prioritisation** — a new user should get summaries
+The rationale for both paths is **onboarding prioritisation under an honest, bounded load** — a
-of their most recent, relevant videos quickly while the older back-catalogue fills in behind,
+new user gets summaries of their most recent videos automatically, while the older back-catalogue
-all within the honest shared rate limit.
+is listed but summarised only on demand, so it never re-drives the shared rate gate every cycle.
 
 ### Newest-first batch ordering (ADR-018)
 
 Within each scheduled pass, `RunOnce` uses a three-phase structure:
 
 1. **Discover + persist**: walk all channels, `UpsertVideo` every candidate (so it appears in
-   the list), apply pre-filters (seen/manual/backoff), collect surviving candidates.
+   the list), apply pre-filters (seen/manual/backoff/**recency**), collect surviving candidates.
+   The recency pre-filter (ADR-020) drops auto-mode videos published before
+   `now - TAPIR_AUTO_SUMMARIZE_WINDOW` unless they are explicitly requested; an undated video is
+   never aged out. They remain persisted/listed — only auto-summarisation is skipped.
 2. **Sort**: order candidates `published_at DESC, NULLS LAST, discovery_pos ASC`. Videos with
    no publish date (schema 001: nullable) sort after all dated content. The sort is in-memory
    (`slices.SortStableFunc`) — at current scale this is fine.
@@ -235,7 +247,8 @@ Within each scheduled pass, `RunOnce` uses a three-phase structure:
 Before (per-channel inline): `[chanA-old, chanA-mid, chanB-new, chanB-null]`
 After  (newest-first):       `[chanB-new, chanA-mid, chanA-old, chanB-null]`
 
-The set of processed videos is identical; only the order within a pass changes.
+The set of *processed* videos now also excludes auto-mode back-catalogue beyond the recency
+window (those stay listed, summarised on demand); within the processed set, only order changes.
 
 ---
 

docs/data-model.md

@@ -10,12 +10,15 @@ only opaque references to them; the secret material lives in ESO/1Password (ADR-
 
 ## Design decisions baked into this model
 
-- **Per-user isolation, not a shared global video table.** The earlier draft proposed a
+- **Per-user isolation for everything except transcripts.** The earlier draft proposed a
-  global `videos`/`transcripts` table deduped across tenants. Rejected for Future B: it
+  global `videos`/`transcripts` table deduped across tenants. **Videos** stay per-user and
-  reintroduces exactly the cross-domain coupling the homelab architecture review is
+  RLS-scoped — a shared video table reintroduces exactly the cross-domain coupling the homelab
-  removing, and at 1–5 users the cost of occasionally re-summarizing the same video is
+  architecture review is removing. **Transcripts**, however, are now shared (ADR-021): keyed by
-  trivial compared to the isolation it would cost. Each user's data is self-contained.
+  `(provider, provider_video_id)`, no `user_id`, **not** RLS-scoped. The cost avoided there is
-  (Revisit only if Future C makes GPU/transcription cost dominate — a new ADR, not a default.)
+  not LLM re-summarization but a rate-gated, reputation-risky caption fetch (ADR-010/014), which
+  is paid per re-fetch regardless of user count — so persisting public caption content once and
+  sharing it strictly beats the coupling it removes. Everything else each user owns is
+  self-contained; `rls_test.go` proves transcripts is the single exception.
 - **Secrets by reference only.** Tables hold a `secret_ref` (opaque string/UUID resolved via
   the `SecretStore` port), never tokens or keys.
 - **The brain sink is just a delivery target.** No brain-specific tables. Whether a summary
@@ -34,7 +37,7 @@ erDiagram
     USER ||--o{ AI_CREDENTIAL : "has (planned)"
     VIDEO_CONNECTION ||--o{ SUBSCRIPTION : "exposes (planned)"
     SUBSCRIPTION ||--o{ VIDEO : "produces (per user)"
-    VIDEO ||--o| TRANSCRIPT : "has at most one"
+    VIDEO }o--o| TRANSCRIPT : "shares one by (provider, provider_video_id) — not FK (ADR-021)"
     VIDEO ||--o| SUMMARY : "has at most one"
     SUMMARY ||--o{ SINK_DELIVERY : "delivered via"
     USER ||--o{ CHANNEL_ERROR : "reports unavailable channels"
@@ -92,12 +95,12 @@ erDiagram
         timestamptz rate_limited_at "backoff clock for 429 retries (migration 007)"
     }
     TRANSCRIPT {
-        uuid video_id PK_FK
+        text provider PK "part of shared key (ADR-021)"
-        uuid user_id FK
+        text provider_video_id PK "part of shared key — the cross-user dedup key"
         text source "captions | none"
         text language
         text content "null when source = none"
-        timestamptz resolved_at
+        timestamptz fetched_at
     }
     SUMMARY {
         uuid id PK
@@ -147,9 +150,10 @@ mechanism.
 
 - **USER** — one row per registered user (Stage 1, ADR-012; no longer single-row). The Tapir-side
   profile; the Dex identity is held separately in `USER_IDENTITY`, not on this row. `auto_summarize`
-  (migration 006) is the per-user mode flag: `TRUE` = auto-summarize every new video. Default is
+  (migration 006) is the per-user mode flag: `TRUE` = auto-summarize new videos **published within
-  **true** for new users (migration 011, ADR-018); existing rows were back-filled via migration 012
+  the recency window** (`TAPIR_AUTO_SUMMARIZE_WINDOW`, default ~7d, ADR-020); older videos are
-  with RLS bypass.
+  listed but summarised on demand. Default is **true** for new users (migration 011, ADR-018);
+  existing rows were back-filled via migration 012 with RLS bypass.
 - **USER_IDENTITY** (migration 004) — the `dex_subject → user_id` map. `dex_subject` is the PK,
   `user_id` a `UNIQUE` FK to `users` with `ON DELETE CASCADE`. This is the bridge resolved at login
   *before* a `user_id` is known, so it is **deliberately not RLS-enabled** (it holds no user data;
@@ -172,8 +176,12 @@ mechanism.
   `transcript_status` and `rate_limited_at` (migration 007) track caption-fetch outcomes for
   rate-limit backoff: `NULL` = not attempted; `rate_limited` = 429 seen, skip until
   `NOW() - rate_limited_at > TAPIR_FETCH_BACKOFF`; `fetched` = resolved; `none` = no transcript.
-- **TRANSCRIPT** — at most one per video. `source = none` records "checked, no usable
+- **TRANSCRIPT** — shared public caption content, one row per `(provider, provider_video_id)`,
-  transcript" so the watcher doesn't reprocess (ADR-007). `content` null in that case.
+  **not** RLS-scoped and carrying no `user_id` (ADR-021). Two users who watch the same video
+  share the one row; the summarize path reads it before any caption fetch, so re-analysis never
+  re-touches YouTube (ADR-010/014). `source = none` records "checked, no usable transcript" so
+  no one reprocesses (ADR-007); `content` null in that case. A transient 429 is never stored
+  here — it stays a per-user retry via `VIDEO.transcript_status`.
 - **SUMMARY** — at most one per video. `fallback_used` + `ai_provider`/`ai_model` make the
   "is local good enough?" question queryable (the Stage 0 quality signal). `highlights`/
   `takeaways` as jsonb to stay schema-flexible while the output format settles.
@@ -230,7 +238,8 @@ queue, doesn't replace it). Deferred until there's a reason.
 
 ## Explicitly out of scope (Future C)
 
-- Global cross-tenant video/transcript dedup (rejected above).
+- Global cross-tenant *video* dedup (rejected above). Note: cross-tenant *transcript* sharing
+  is now in scope and shipped (ADR-021); only the videos half stays per-user.
 - Sharding / per-tenant physical databases.
 - Soft-delete + full audit trail on connections/credentials (a Stage 2 hardening item; add
   via ADR when Stage 2 work starts).

docs/homelab-integration.md

@@ -27,6 +27,18 @@ it** — endpoints and aliases drift, and this file is a snapshot (2026-06-06),
   `iguana/deepseek-r1-14b`) is preferred for summary quality if its latency/output is acceptable.
   The `max_tokens` fix below means thinking models no longer return empty content, so they are now
   viable choices, not blocked ones. Do not assume a coder alias is right for prose.
+- **Summarizer fallback chain (ADR-022).** The primary alias is the *first* of an ordered chain;
+  on failure or unparseable output the summarizer advances to the next model. All reached through
+  the same gateway by alias.
+  - `TAPIR_FALLBACK_MODEL` — local fallback. **Default `koala/phi4-14b`.** Empty disables it.
+  - `TAPIR_CLOUD_FALLBACK_MODEL` — worst-case EXTERNAL fallback. **Default `berget/mistral-small`.**
+    **Set this empty (`""`) for any client/NDA deployment** so content never leaves the local
+    stack — the chain then contains only local endpoints.
+  - `TAPIR_SUMMARY_MAX_TOKENS` — per-summary completion budget. **Default `1500`.** Small on
+    purpose: with the old 8192 budget, prompt + completion overflowed `phi4-mini`'s 8k window.
+  - `TAPIR_MAX_TRANSCRIPT_CHARS` — transcript truncation budget sent to the model. **Default
+    `18000`** (~fits an 8k-context model). `0` disables truncation. Prevents the context-overflow
+    HTTP 400 a long transcript caused on `phi4-mini`.
 - **Thinking models need an explicit `max_tokens`.** qwen3 / deepseek-r1 spend the budget on
   reasoning and return **empty content** if `max_tokens` is too low (or unset). The summarizer's
   parser treats an empty summary as an error for exactly this reason. **Done (2026-06-02, Worker F):**

docs/specs/newest-first-ordering.md

@@ -1,5 +1,11 @@
 # Spec — Newest-first batch ordering + honest "Try now" / prioritisation docs
 
+> **Extended by ADR-020 (2026-06-08).** This spec covers the *batch processing* order within a
+> pass. ADR-020 adds (a) a recency pre-filter — auto mode skips videos published before
+> `TAPIR_AUTO_SUMMARIZE_WINDOW`, listed but summarised on demand — and (b) the same
+> `published_at DESC NULLS LAST` ordering on the **list read** (`ListVideos`), which previously
+> sorted by `seen_at`. See `DECISIONS.md` ADR-020.
+
 **Repo:** tapir · **Size:** small · **Solo session.**
 
 **Why.** Product intent (maintainer, 2026-06-06): a new user should get summaries of their

docs/specs/scheduled-discovery.md

@@ -1,5 +1,10 @@
 # Spec — In-process scheduled discovery + auto-summarize + rate-gate finish
 
+> **Extended by ADR-020 (2026-06-08).** Auto-summarize is no longer "every unseen video": the
+> scheduler now skips videos published before `TAPIR_AUTO_SUMMARIZE_WINDOW` (default ~7d) unless
+> explicitly requested, so a back-catalogue does not re-drive the rate gate every cycle. See
+> `DECISIONS.md` ADR-020.
+
 **Repo:** tapir · **Size:** medium · **Solo session** (not a swarm).
 
 **Why this exists.** The Stage-0 gate ("me or a friend returns and reads/acts in ≥2 separate

docs/specs/video-card-states.md

@@ -1,5 +1,12 @@
 # Spec — Unify video-card states: one "Summarize now" verb, honest no-captions state
 
+> **Superseded in part by ADR-020 (2026-06-08).** The five card states still hold, but the copy
+> changed: the nudge verb is now **"Summarize"** (not "Summarize now"), the rate-limited state
+> reads **"In queue"** (not "Fetching soon…"), and the queued state reads **"summarizing
+> shortly"** (not "waiting for the next run"). The list also now collapses older un-summarized
+> and caption-less videos. See `DECISIONS.md` ADR-020 and `views.templ` (`VideoCard`) for the
+> current copy; this doc is kept as the original design record.
+
 **Repo:** tapir · **Size:** small, **view-layer only** (`views.templ` + a little CSS in
 `view.go`; regenerate `views_templ.go`). No handler, store, or DB change. The two existing
 handlers (`/summarize`, `/retry-now`) stay exactly as they are — only what the card *shows*

docs/ui-spec.md

@@ -179,3 +179,4 @@ distinguishable.
 | **"Summarize now" foreground path** | Unified quiet nudge button on actionable non-summarized cards. Five explicit card states — (1) summarized: chip + no button; (2) no captions (`transcript_status = 'none'`): "No transcript available", no button; (3) queued: "Queued" chip, no button; (4) rate-limited: "Fetching soon…" + "Summarize now" → `POST /v/{id}/retry-now` (clears `rate_limited_at`, triggers engine); (5) pending: "Not summarized" + "Summarize now" → `POST /v/{id}/summarize` (queues + triggers engine). One verb, one style (`.btn-quiet`); backend difference invisible to user. Both handlers call `ProcessVideo` through `globalFetchGate`. Rate gate respected, not bypassed — this is onboarding prioritisation. | Fast onboarding value; honest dead-end for no-captions videos (no button that fails). | `internal/web/handlers.go` (`handleRetryNow`, `handleRequestSummarize`); `internal/web/views.templ` (`VideoCard`) |
 | **Pipeline stats bar** | A one-line status bar above the video list: `N summarized · M fetching soon · K no captions`. Computed from the unfiltered row set; hidden when all videos are summarized. Gives the user a clear read on pipeline state without any interaction. | Replaces the "why is nothing happening?" confusion when most videos are pending or rate-limited. | `internal/web/view.go` (`PipelineStats`, `pipelineStats`) |
 | **Unavailable channels (account page)** | The `/account` page shows a "Unavailable channels" section when any channels returned HTTP 404 on the last discovery pass. Lists channel name, an "unavailable" badge, and the first-seen date. Data sourced from the `channel_errors` table (migration 013). | Surfaces silent failures so users know why some subscribed channels produce no new videos. | migration 013; `internal/web/account.go`; `internal/adapters/youtube/youtube.go` (`domain.ErrChannelUnavailable`) |
+| **Recency window + sparse-state honesty (ADR-020)** | Supersedes the copy/sort in the rows above. Auto-summarize is bounded to videos published within `TAPIR_AUTO_SUMMARIZE_WINDOW` (~7d); older un-summarized videos collapse behind a single "Show N older videos — summarize on demand" disclosure, and caption-less videos collapse to a one-line count (not N cards). List order is now `summarized-first, published_at DESC NULLS LAST`. Copy reframed for honest scarcity: pipeline bar reads "N ready · M in queue · K no captions" (no "fetching soon"); a gradual-fill note explains the rate limit; the nudge verb is "Summarize" (not "Summarize now"); the queued card says "summarizing shortly"; the empty-connected state drops the impossible `tapir run` instruction. Detail leads with Takeaways. Filters slimmed (no date pickers; hidden when empty); watched/skipped segmented; back link on detail; empty terms checkbox removed. | Make the sparse reality legible and honest instead of implying abundance/imminence; bound auto load so the back-catalogue doesn't re-drive the caption gate. Never fetch harder — scarcity is surfaced, not engineered around. | ADR-020; `2384c47`, `3df0459`, `40b703e`, `a1a5217`, `4a0a56e`, `9bf1c31`, `980638d`, `12fb031`, `f775441`, `51aa5d9` |

docs/use-cases/ai_routing.feature

@@ -31,5 +31,17 @@ Feature: Local-first AI with optional BYO fallback
     When any transcript is summarized
     Then my content is only ever sent to the local AI stack
 
-  # "Reliably" is operationalized as: Primary returned without error within timeout.
+  Scenario: A model returns unparseable output and the next endpoint succeeds
+    Given the local AI stack is available
+    But the primary model returns output that cannot be parsed into a summary
+    And a fallback model is configured
+    When a transcript is summarized
+    Then Tapir falls back to the next model in the chain
+    And the summary records fallback_used as true
+
+  # "Reliably" is operationalized as: an endpoint returned a PARSEABLE summary
+  # within timeout. A 200 with malformed JSON (or highlights emitted as a bare
+  # string) counts as a failure and advances the chain (ADR-022). Endpoints are
+  # tried in order, locals first, so the external worst-case model only ever sees
+  # content after every local endpoint has failed.
   # Quality scoring may be added later without changing these scenarios.

docs/use-cases/connect_account.feature

@@ -10,12 +10,26 @@ Feature: Connect and manage video accounts
     And my refresh token is stored only as a secret reference
     And my subscriptions are synced
 
+  @pending
+  # Vimeo connect is not built yet (provider label exists; no connect flow or test).
   Scenario: Connect a Vimeo account
     Given I have no connected video accounts
     When I connect my Vimeo account
     Then the connection is stored with status "active"
     And my subscriptions are synced
 
+  Scenario: Connecting an account discovers videos immediately
+    Given I have no connected video accounts
+    When I connect my YouTube account
+    Then a discovery pass for my account is triggered right away
+    And I do not have to wait for the next scheduled pass to see my videos
+
+  Scenario: Connecting summarizes my newest videos right away
+    Given I have no connected video accounts
+    When I connect my YouTube account
+    Then up to the onboarding cap of my newest videos are summarized through the rate gate
+    And the rest are left to the scheduled recency-bounded pass
+
   Scenario: Tokens are never stored in the clear
     When I connect any video account
     Then no OAuth token value is stored in the database
@@ -28,6 +42,9 @@ Feature: Connect and manage video accounts
     And no new videos are watched for that connection
     And my existing summaries remain readable
 
+  @pending
+  # Per-provider BYO credential config is not built as a web flow yet (the summarizer
+  # supports a fallback endpoint, but there is no user-facing BYO setup + its test).
   Scenario Outline: BYO AI credential is optional and per-provider
     When I configure a BYO provider "<provider>"
     Then the credential is stored only as a secret reference

docs/use-cases/landing_page.feature

@@ -19,6 +19,9 @@ Feature: Public landing page
     Then I see a link to my summaries
     And I see a way to log out
 
+  @pending
+  # Behaviour ships (logout redirects to /welcome) but is not unit-tested: logout lives in
+  # the OIDC Auth impl and StubAuth has no routes to exercise it cheaply.
   Scenario: Logging out returns to the welcome page
     Given I am logged in
     When I log out

docs/use-cases/paste_url.feature

@@ -0,0 +1,30 @@
+Feature: Paste a YouTube URL to summarize any video
+  As a user
+  I want to paste a YouTube link and get a summary
+  So that I can pull the specific video I want now, even from channels I don't follow
+
+  Scenario: Paste a valid YouTube URL
+    Given I am connected
+    When I paste a valid YouTube video URL
+    Then the video is added to my feed scoped to me
+    And it is queued for summarization through the shared rate gate
+
+  Scenario: Pasting an invalid link is rejected
+    When I paste something that is not a YouTube video URL
+    Then I get a clear error and nothing is added
+
+  Scenario: Pasting a video that cannot be found is honest
+    When I paste a URL whose video cannot be found
+    Then I am told it couldn't be found and nothing is added
+
+  Scenario: Pasting the same video twice does not duplicate it
+    Given I have pasted a video
+    When I paste the same video again
+    Then my feed still has exactly one entry for it
+
+  @pending
+  # Covered by the engine's ADR-010 no-transcript terminal state (degrade-never-error);
+  # there is no paste-specific test for it.
+  Scenario: A pasted video with no captions resolves honestly
+    When I paste a video that has no captions
+    Then it resolves to the "no transcript available" terminal state

docs/use-cases/registration.feature

@@ -34,6 +34,9 @@ Feature: Register and manage a multi-user account
     And the other user's data remains intact
     And my Dex identity is left intact
 
+  @pending
+  # Re-registration after delete is supported by design (delete leaves the Dex identity,
+  # ADR-013) but has no dedicated end-to-end test yet.
   Scenario: A deleted user can register again as a fresh account
     Given I deleted my Tapir account but my Dex identity still exists
     When I sign in again

docs/use-cases/summarize_mode.feature

@@ -6,15 +6,25 @@ Feature: Choose how new videos get summarized
   Background:
     Given I am a registered user with a connected video account
 
-  Scenario: Auto mode summarizes every new video
+  Scenario: Auto mode summarizes recent new videos automatically
     Given my summarization mode is "auto"
-    When a subscribed channel posts a new video with captions
+    When a subscribed channel posts a new video with captions within the recency window
     Then Tapir summarizes it without my asking
     And the summary appears in my list
 
-  Scenario: Manual mode is the default and leaves new videos unsummarized
+  Scenario: Auto mode lists older videos without summarizing them
-    Given I have not changed my summarization mode
+    Given my summarization mode is "auto"
-    Then my mode is "manual"
+    When discovery finds a video published before the recency window
+    Then the video appears in my list with no summary
+    And it is not summarized automatically
+    And I can still summarize it on demand with "Summarize"
+
+  Scenario: Automatic is the default for a new user
+    Given I have just registered
+    Then my summarization mode is "auto"
+
+  Scenario: Manual mode leaves new videos unsummarized
+    Given my summarization mode is "manual"
     When a subscribed channel posts a new video with captions
     Then the video appears in my list with no summary
     And nothing is summarized until I request it
@@ -30,3 +40,8 @@ Feature: Choose how new videos get summarized
   # auto_summarize is a per-user setting and summarize_requested is a per-video queue
   # flag (migration 006). The web button sets the flag; `tapir run` processes both the
   # auto videos and the manually queued ones, then clears the flag.
+  #
+  # Recency bound (ADR-020): in auto mode the scheduler only summarizes videos published
+  # within TAPIR_AUTO_SUMMARIZE_WINDOW (default ~7d); older videos are discovered and
+  # listed but wait for an explicit "Summarize" — so a back-catalogue does not re-drive
+  # the per-IP caption gate (ADR-014) every cycle. A manual request bypasses the bound.

docs/use-cases/summarize_new_video.feature

@@ -31,5 +31,16 @@ Feature: Summarize new videos from subscribed channels
     When the watcher sees "Designing for Attention" again
     Then Tapir does not produce a second summary for it
 
+  Scenario: Re-analyzing a stored video does not re-fetch its transcript
+    Given a transcript for "Designing for Attention" is already stored
+    When the video is summarized again
+    Then Tapir reads the stored transcript
+    And Tapir does not fetch captions from YouTube
+
   # Captions-first is the core path (ADR-007). Audio-download + speech-to-text is
   # deferred and intentionally has no scenario here yet.
+  #
+  # Transcript persistence (ADR-021): the stored transcript is shared, keyed by
+  # (provider, provider_video_id) and read before any caption fetch, so the
+  # re-analysis scenario above also covers paste-a-URL and the onboarding burst —
+  # both summarize through the same engine chokepoint.

internal/adapters/llm/client.go

@@ -34,15 +34,35 @@ type Client struct {
 	httpClient *http.Client
 }
 
+// Option configures a Client at construction. Variadic so the existing 4-arg
+// call sites stay valid as new knobs are added.
+type Option func(*Client)
+
+// WithMaxTokens overrides the per-request completion budget. The summarizer uses
+// this to cap completion for small-context models (e.g. koala/phi4-mini, 8k):
+// with the default 8192 budget, prompt + max_tokens overflows an 8k context and
+// the gateway returns HTTP 400. A non-positive n is ignored (keeps the default).
+func WithMaxTokens(n int) Option {
+	return func(c *Client) {
+		if n > 0 {
+			c.maxTokens = n
+		}
+	}
+}
+
 // New constructs a Client.
-func New(baseURL, apiKey, model string, timeout time.Duration) *Client {
+func New(baseURL, apiKey, model string, timeout time.Duration, opts ...Option) *Client {
-	return &Client{
+	c := &Client{
 		baseURL:    strings.TrimRight(baseURL, "/"),
 		apiKey:     apiKey,
 		model:      model,
 		maxTokens:  defaultMaxTokens,
 		httpClient: &http.Client{Timeout: timeout},
 	}
+	for _, opt := range opts {
+		opt(c)
+	}
+	return c
 }
 
 type chatRequest struct {

internal/adapters/llm/client_test.go

@@ -64,6 +64,27 @@ func TestClient_SendsMaxTokens(t *testing.T) {
 	}
 }
 
+// TestClient_WithMaxTokens overrides the completion budget — the summarizer caps
+// it small so prompt + max_tokens fits a small-context model's window (8k).
+func TestClient_WithMaxTokens(t *testing.T) {
+	var body chatRequest
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		_ = json.NewDecoder(r.Body).Decode(&body)
+		_ = json.NewEncoder(w).Encode(map[string]any{
+			"choices": []map[string]any{{"message": map[string]any{"content": "ok"}}},
+		})
+	}))
+	defer srv.Close()
+
+	c := New(srv.URL, "", "test-model", 10*time.Second, WithMaxTokens(1500))
+	if _, err := c.Complete(context.Background(), "sys", "user"); err != nil {
+		t.Fatalf("Complete: %v", err)
+	}
+	if body.MaxTokens != 1500 {
+		t.Errorf("max_tokens = %d, want 1500", body.MaxTokens)
+	}
+}
+
 func TestClient_ReturnsErrorOnNon200(t *testing.T) {
 	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		http.Error(w, "overloaded", http.StatusServiceUnavailable)

internal/adapters/store/account.go

@@ -10,10 +10,13 @@ import (
 // DeleteUser permanently removes a user and all of their data. It runs through
 // withUser so RLS confines every statement to the calling user's own rows.
 //
-// Deleting the users row cascades (ON DELETE CASCADE) to videos, transcripts,
+// Deleting the users row cascades (ON DELETE CASCADE) to videos, summaries
-// summaries (→ sink_deliveries), video_connections, and the user_identities map
+// (→ sink_deliveries), video_connections, and the user_identities map —
-// — referential-integrity cascades bypass RLS, so a user's child rows are removed
+// referential-integrity cascades bypass RLS, so a user's child rows are removed
-// even though the deleting connection is scoped. summary_actions and login_events
+// even though the deleting connection is scoped. Transcripts are NOT removed:
+// since ADR-021 they are shared public content keyed by (provider,
+// provider_video_id) with no user_id, so another user may still reference the
+// same row — a user deletion must not strip shared caption content. summary_actions and login_events
 // are the exceptions: each carries a user_id but has NO foreign key to users
 // (migrations 002 and 010), so the cascade does not reach them; they are deleted
 // explicitly in the same scoped transaction. Deleting an absent user is a no-op

internal/adapters/store/migrate_test.go

@@ -53,7 +53,11 @@ func TestMigration010LoginEventsUpDown(t *testing.T) {
 	require.True(t, loginEventsExists(t), "login_events must exist at latest migration")
 
 	m := fileMigrator(t)
-	// 011, 012, 013 sit above 010; step them down first so 010 is exercised in isolation.
+	// 011..015 sit above 010; step them down first so 010 is exercised in isolation.
+	require.NoError(t, m.Steps(-1), "down 015 reshapes transcripts, login_events intact")
+	require.True(t, loginEventsExists(t), "015 down leaves login_events intact")
+	require.NoError(t, m.Steps(-1), "down 014 drops channel_title, login_events intact")
+	require.True(t, loginEventsExists(t), "014 down leaves login_events intact")
 	require.NoError(t, m.Steps(-1), "down 013 drops channel_errors, login_events intact")
 	require.True(t, loginEventsExists(t), "013 down leaves login_events intact")
 	require.NoError(t, m.Steps(-1), "down 012 is a no-op, login_events intact")
@@ -64,7 +68,7 @@ func TestMigration010LoginEventsUpDown(t *testing.T) {
 	require.NoError(t, m.Steps(-1), "down 010 must drop login_events")
 	require.False(t, loginEventsExists(t), "login_events must be gone after the down migration")
 
-	require.NoError(t, m.Steps(4), "up must recreate 010 then re-apply 011, 012, 013")
+	require.NoError(t, m.Steps(6), "up must recreate 010 then re-apply 011..015")
 	require.True(t, loginEventsExists(t), "login_events must be restored after the up migration")
 }
 
@@ -87,6 +91,8 @@ func TestMigration011AutoSummarizeDefaultUpDown(t *testing.T) {
 	require.Equal(t, "true", autoSummarizeDefault(t), "011 sets the default to TRUE")
 
 	m := fileMigrator(t)
+	require.NoError(t, m.Steps(-1), "down 015 reshapes transcripts")
+	require.NoError(t, m.Steps(-1), "down 014 drops channel_title")
 	require.NoError(t, m.Steps(-1), "down 013 drops channel_errors")
 	require.NoError(t, m.Steps(-1), "down 012 is a no-op")
 	require.NoError(t, m.Steps(-1), "down 011 reverts the column default")
@@ -96,6 +102,35 @@ func TestMigration011AutoSummarizeDefaultUpDown(t *testing.T) {
 	require.Equal(t, "true", autoSummarizeDefault(t))
 	require.NoError(t, m.Steps(1), "up 012 runs clean (no FORCE RLS on fresh schema)")
 	require.NoError(t, m.Steps(1), "up 013 creates channel_errors")
+	require.NoError(t, m.Steps(1), "up 014 recreates channel_title")
+	require.NoError(t, m.Steps(1), "up 015 reshapes transcripts to shared")
+}
+
+// channelTitleExists reports whether videos.channel_title is present.
+func channelTitleExists(t *testing.T) bool {
+	t.Helper()
+	var exists bool
+	require.NoError(t, rawPool(t).QueryRow(context.Background(),
+		`SELECT EXISTS (SELECT 1 FROM information_schema.columns
+		  WHERE table_name = 'videos' AND column_name = 'channel_title')`).Scan(&exists))
+	return exists
+}
+
+// TestMigration014VideoChannelTitleUpDown proves 014 is reversible: down drops
+// videos.channel_title, up recreates it.
+func TestMigration014VideoChannelTitleUpDown(t *testing.T) {
+	newStore(t) // latest (014 applied)
+	require.True(t, channelTitleExists(t), "channel_title exists at latest migration")
+
+	m := fileMigrator(t)
+	require.NoError(t, m.Steps(-1), "down 015 reshapes transcripts, channel_title intact")
+	require.True(t, channelTitleExists(t), "015 down leaves channel_title intact")
+	require.NoError(t, m.Steps(-1), "down 014 must drop channel_title")
+	require.False(t, channelTitleExists(t), "channel_title must be gone after the down migration")
+
+	require.NoError(t, m.Steps(1), "up 014 must recreate channel_title")
+	require.True(t, channelTitleExists(t), "channel_title must be restored after the up migration")
+	require.NoError(t, m.Steps(1), "up 015 restores the shared transcripts shape (HEAD)")
 }
 
 // TestMigration012FixAutoSummarizeRLS proves 012 runs cleanly and flips any

internal/adapters/store/migrations/014_video_channel_title.down.sql

@@ -0,0 +1 @@
+ALTER TABLE videos DROP COLUMN channel_title;

internal/adapters/store/migrations/014_video_channel_title.up.sql

@@ -0,0 +1,8 @@
+-- Store the source channel's title per video so the list can offer a real
+-- channel filter (multi-select of the user's channels) instead of the dead
+-- free-text field that only ever matched the provider string. Nullable: existing
+-- rows backfill on the next discovery pass (UpsertVideo writes it); pasted videos
+-- get it immediately from videos.list. No FK to a channels table at Stage 0 — the
+-- title is a denormalised display/filter value, consistent with the existing
+-- subscription_id-stays-NULL stance (data-model.md).
+ALTER TABLE videos ADD COLUMN channel_title TEXT;

internal/adapters/store/migrations/015_shared_transcripts.down.sql

@@ -0,0 +1,19 @@
+-- Down 015: restore the per-user RLS-scoped transcripts shape (001 + 003).
+DROP TABLE transcripts;
+
+CREATE TABLE transcripts (
+    video_id    UUID PRIMARY KEY REFERENCES videos(id) ON DELETE CASCADE,
+    user_id     UUID NOT NULL REFERENCES users(id) ON DELETE CASCADE,
+    source      TEXT NOT NULL,
+    language    TEXT,
+    content     TEXT,
+    resolved_at TIMESTAMPTZ NOT NULL DEFAULT NOW()
+);
+
+CREATE INDEX idx_transcripts_user_id ON transcripts(user_id);
+
+ALTER TABLE transcripts ENABLE ROW LEVEL SECURITY;
+ALTER TABLE transcripts FORCE ROW LEVEL SECURITY;
+CREATE POLICY transcripts_isolation ON transcripts
+    FOR ALL
+    USING (user_id = current_setting('tapir.current_user_id', true)::uuid);

internal/adapters/store/migrations/015_shared_transcripts.up.sql

@@ -0,0 +1,30 @@
+-- Migration 015: transcripts become SHARED public-content storage (ADR-021).
+--
+-- The per-user transcripts table from 001 (PK videos.id, user_id NOT NULL, RLS
+-- FORCEd in 003) was dead: no application code ever read or wrote it — only the
+-- transcript_status columns on `videos` (007) carried fetch outcomes. ADR-021
+-- repurposes it as the single shared store of public caption content, keyed by
+-- the cross-user dedup key (provider, provider_video_id) — the video's public
+-- identity, not Tapir's per-user videos.id — so re-analysis never re-fetches
+-- from YouTube (ADR-010/014).
+--
+-- It holds ONLY public caption content + the video's public id (nothing
+-- user-identifying), so it is deliberately NOT RLS-scoped: no user_id, no
+-- policy, no FORCE. This is the single, intentional exception to the ADR-012
+-- isolation boundary; rls_test.go asserts the boundary is exactly here and
+-- nowhere else. Dropping the old table drops its RLS policy with it; it held no
+-- real data, so drop+recreate loses nothing.
+DROP TABLE transcripts;
+
+CREATE TABLE transcripts (
+    provider          TEXT NOT NULL,
+    provider_video_id TEXT NOT NULL,
+    source            TEXT NOT NULL, -- 'captions' (content set) | 'none' (no captions; content NULL)
+    language          TEXT,
+    content           TEXT,
+    fetched_at        TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    PRIMARY KEY (provider, provider_video_id)
+);
+
+COMMENT ON TABLE transcripts IS
+    'Shared public caption content keyed by (provider, provider_video_id). NOT RLS-scoped — public content only, de-facto cross-user dedup (ADR-021).';

internal/adapters/store/reads.go

@@ -29,6 +29,7 @@ type SummaryRow struct {
 	ProviderVideoID string    // videos.provider_video_id; empty when no videos row
 	Title           string    // videos.title; empty when no videos row
 	Channel         string    // videos.provider for now; empty when no videos row
+	ChannelTitle    string    // videos.channel_title; the source channel, for display + filtering
 	URL             string    // videos.url; empty when no videos row
 	PublishedAt     time.Time // videos.published_at; zero when absent
 	Summary         string
@@ -137,7 +138,8 @@ const selectVideo = `
 	       COALESCE(s.created_at, v.seen_at),
 	       (s.id IS NOT NULL) AS summarized,
 	       v.summarize_requested,
-	       COALESCE(v.transcript_status, '')
+	       COALESCE(v.transcript_status, ''),
+	       COALESCE(v.channel_title, '')
 	  FROM videos v
 	  LEFT JOIN summaries s ON s.video_id = v.id AND s.user_id = v.user_id`
 
@@ -254,6 +256,7 @@ func scanVideoRow(rows pgx.Row) (SummaryRow, error) {
 		&row.Summarized,
 		&row.SummarizeRequested,
 		&row.TranscriptStatus,
+		&row.ChannelTitle,
 	); err != nil {
 		return SummaryRow{}, fmt.Errorf("store: scan video: %w", err)
 	}

internal/adapters/store/rls_test.go

@@ -22,9 +22,11 @@ import (
 // (no GUC set → zero rows) proves the enforcement path is live, not bypassed.
 
 // userIsolatedTables are the tables that carry a user_id and whose policy keys
-// directly off the tapir.current_user_id GUC.
+// directly off the tapir.current_user_id GUC. transcripts is deliberately ABSENT
+// — ADR-021 made it shared public content (non-RLS); TestTranscriptsTableIsSharedNotRLS
+// proves that is the only place the isolation boundary moved.
 var userIsolatedTables = []string{
-	"users", "videos", "transcripts", "summaries", "summary_actions", "login_events", "video_connections",
+	"users", "videos", "summaries", "summary_actions", "login_events", "video_connections",
 }
 
 // allIsolatedTables adds sink_deliveries, whose ownership is derived from its
@@ -38,9 +40,10 @@ type seeded struct {
 	summaryID string
 }
 
-// seedUser inserts one full chain (user → video → transcript → summary →
+// seedUser inserts one full chain (user → video → summary → action → delivery)
-// action → delivery) as the superuser pool, which bypasses RLS so both users'
+// as the superuser pool, which bypasses RLS so both users' data lands regardless
-// data lands regardless of the GUC.
+// of the GUC. Transcripts are NOT seeded here: they are shared, non-RLS public
+// content (ADR-021), so they have no place in a per-user isolation chain.
 func seedUser(t *testing.T, p *pgxpool.Pool, userID string) seeded {
 	t.Helper()
 	ctx := context.Background()
@@ -54,11 +57,6 @@ func seedUser(t *testing.T, p *pgxpool.Pool, userID string) seeded {
 		 VALUES ($1, 'youtube', $2, 'title') RETURNING id`,
 		userID, "vid-"+userID).Scan(&videoID))
 
-	_, err = p.Exec(ctx,
-		`INSERT INTO transcripts (video_id, user_id, source, content)
-		 VALUES ($1, $2, 'captions', 'words')`, videoID, userID)
-	require.NoError(t, err)
-
 	var summaryID string
 	require.NoError(t, p.QueryRow(ctx,
 		`INSERT INTO summaries (user_id, video_id, summary) VALUES ($1, $2, 'sum')
@@ -92,9 +90,16 @@ func appPool(t *testing.T, super *pgxpool.Pool) *pgxpool.Pool {
 	t.Helper()
 	ctx := context.Background()
 
-	// Idempotent across test runs (schema/role persist for the TestMain PG).
+	// Idempotent across tests AND runs: the role persists for the TestMain PG and
-	_, _ = super.Exec(ctx, `DROP ROLE IF EXISTS app`)
+	// owns granted privileges, so a plain DROP ROLE fails once any GRANT exists
-	_, err := super.Exec(ctx, `CREATE ROLE app LOGIN PASSWORD 'app'`)
+	// (and more than one test now builds an app pool). Create only if absent; the
+	// GRANTs below are themselves idempotent.
+	_, err := super.Exec(ctx,
+		`DO $$ BEGIN
+		    IF NOT EXISTS (SELECT FROM pg_roles WHERE rolname = 'app') THEN
+		        CREATE ROLE app LOGIN PASSWORD 'app';
+		    END IF;
+		 END $$`)
 	require.NoError(t, err)
 	_, err = super.Exec(ctx, `GRANT USAGE ON SCHEMA public TO app`)
 	require.NoError(t, err)
@@ -186,7 +191,6 @@ func TestRLSEnforcesPerUserIsolation(t *testing.T) {
 		{"update users", `UPDATE users SET display_name = 'hacked' WHERE id = $1`, b.userID},
 		{"update videos", `UPDATE videos SET title = 'hacked' WHERE user_id = $1`, b.userID},
 		{"queue videos summarize", `UPDATE videos SET summarize_requested = TRUE WHERE id = $1`, b.videoID},
-		{"update transcripts", `UPDATE transcripts SET content = 'hacked' WHERE user_id = $1`, b.userID},
 		{"update summaries", `UPDATE summaries SET summary = 'hacked' WHERE user_id = $1`, b.userID},
 		{"update summary_actions", `UPDATE summary_actions SET action = 'skipped' WHERE user_id = $1`, b.userID},
 		{"update login_events", `UPDATE login_events SET seen_at = NOW() WHERE user_id = $1`, b.userID},
@@ -235,3 +239,55 @@ func TestRLSEnforcesPerUserIsolation(t *testing.T) {
 
 	_ = a // a's ids are seeded for the symmetric read assertions above
 }
+
+// TestTranscriptsTableIsSharedNotRLS is the ADR-021 isolation proof: transcripts
+// is the ONE shared, non-RLS surface, and the public-content classification
+// leaked to nothing else. It is the inverse of TestRLSEnforcesPerUserIsolation —
+// where that asserts deny-all on every user-owned table, this asserts transcripts
+// is readable and writable with no user scope at all, holds no user_id, and is
+// the single table with row-level security switched off.
+func TestTranscriptsTableIsSharedNotRLS(t *testing.T) {
+	newStore(t)
+	super := rawPool(t)
+	resetDB(t, super)
+	app := appPool(t, super)
+	ctx := context.Background()
+
+	// 1. Shared + non-RLS: with NO GUC set, the app role both writes and reads a
+	//    transcript. On an RLS table this would be deny-all (zero rows), exactly as
+	//    the main isolation test asserts for every user-owned table.
+	_, err := app.Exec(ctx,
+		`INSERT INTO transcripts (provider, provider_video_id, source, content)
+		 VALUES ('youtube', 'shared-vid', 'captions', 'public words')`)
+	require.NoError(t, err, "app role must write shared transcript content with no user scope")
+	require.Equal(t, 1, scopedCount(t, app, "", "transcripts"),
+		"transcripts must be readable with NO user scope — it is shared, non-RLS (ADR-021)")
+
+	// 2. No user_id column: the table holds only public caption content + the
+	//    video's public id, nothing user-identifying.
+	var hasUserID bool
+	require.NoError(t, super.QueryRow(ctx,
+		`SELECT EXISTS (SELECT 1 FROM information_schema.columns
+		                 WHERE table_name = 'transcripts' AND column_name = 'user_id')`).Scan(&hasUserID))
+	require.False(t, hasUserID, "transcripts must carry no user_id (ADR-021 public content)")
+
+	// 3. The boundary is EXACTLY here: every user-owned table still has row-level
+	//    security enabled; transcripts alone has it off. This is the proof the
+	//    non-RLS classification was applied to transcripts and leaked nowhere else.
+	for _, table := range allIsolatedTables {
+		require.True(t, rlsEnabled(t, super, table),
+			"%s must still enforce row-level security — isolation must not have regressed", table)
+	}
+	require.False(t, rlsEnabled(t, super, "transcripts"),
+		"transcripts must be the single table with row-level security OFF (the one shared surface)")
+}
+
+// rlsEnabled reports whether a public table has ROW LEVEL SECURITY enabled.
+func rlsEnabled(t *testing.T, p *pgxpool.Pool, table string) bool {
+	t.Helper()
+	var enabled bool
+	require.NoError(t, p.QueryRow(context.Background(),
+		`SELECT relrowsecurity FROM pg_class
+		  WHERE relname = $1 AND relnamespace = 'public'::regnamespace`, table).Scan(&enabled))
+	return enabled
+}

internal/adapters/store/transcript.go

@@ -0,0 +1,66 @@
+package store
+
+import (
+	"context"
+	"errors"
+	"fmt"
+
+	"github.com/jackc/pgx/v5"
+
+	"gitea.d-ma.be/mathias/tapir/internal/domain"
+)
+
+// GetTranscript returns the shared, stored transcript for a video keyed by the
+// cross-user dedup key (provider, providerVideoID), and whether one exists
+// (ADR-021). It reads via the raw pool, NOT withUser: the table holds public
+// content with no user_id and no RLS policy, so it is shared across users by
+// construction. A stored SourceNone is a real hit (ok == true, HasText() ==
+// false) — a known caption-less video, so the caller skips without re-fetching.
+func (s *Store) GetTranscript(ctx context.Context, provider, providerVideoID string) (domain.Transcript, bool, error) {
+	var source, lang, content string
+	err := s.pool.QueryRow(ctx,
+		`SELECT source, COALESCE(language, ''), COALESCE(content, '')
+		   FROM transcripts WHERE provider = $1 AND provider_video_id = $2`,
+		provider, providerVideoID).Scan(&source, &lang, &content)
+	if errors.Is(err, pgx.ErrNoRows) {
+		return domain.Transcript{}, false, nil
+	}
+	if err != nil {
+		return domain.Transcript{}, false, fmt.Errorf("store: get transcript: %w", err)
+	}
+	return domain.Transcript{
+		Source:   domain.TranscriptSource(source),
+		Language: lang,
+		Content:  content,
+	}, true, nil
+}
+
+// SaveTranscript upserts the shared transcript for (provider, providerVideoID).
+// Only terminal outcomes belong here: SourceCaptions (with text) or SourceNone
+// (no captions). A transient SourceRateLimited is rejected so persistence never
+// masks a 429 as a permanent absence — that stays a per-user retry (ADR-014).
+// Last write wins on conflict (a later re-fetch may correct an entry). It writes
+// via the raw pool, NOT withUser — public content, shared, non-RLS (ADR-021).
+func (s *Store) SaveTranscript(ctx context.Context, provider, providerVideoID string, t domain.Transcript) error {
+	switch t.Source {
+	case domain.SourceCaptions, domain.SourceNone:
+		// terminal — persist
+	case domain.SourceRateLimited:
+		return fmt.Errorf("store: refusing to persist transient rate-limited transcript for %s/%s", provider, providerVideoID)
+	default:
+		return fmt.Errorf("store: invalid transcript source %q", t.Source)
+	}
+	_, err := s.pool.Exec(ctx,
+		`INSERT INTO transcripts (provider, provider_video_id, source, language, content)
+		      VALUES ($1, $2, $3, NULLIF($4, ''), NULLIF($5, ''))
+		 ON CONFLICT (provider, provider_video_id)
+		 DO UPDATE SET source     = EXCLUDED.source,
+		               language   = EXCLUDED.language,
+		               content    = EXCLUDED.content,
+		               fetched_at = NOW()`,
+		provider, providerVideoID, string(t.Source), t.Language, t.Content)
+	if err != nil {
+		return fmt.Errorf("store: save transcript: %w", err)
+	}
+	return nil
+}

internal/adapters/store/transcript_test.go

@@ -0,0 +1,87 @@
+package store_test
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+
+	"gitea.d-ma.be/mathias/tapir/internal/adapters/store"
+	"gitea.d-ma.be/mathias/tapir/internal/domain"
+	"gitea.d-ma.be/mathias/tapir/internal/ports"
+)
+
+// Static check: Store satisfies the shared TranscriptStore port (ADR-021).
+var _ ports.TranscriptStore = (*store.Store)(nil)
+
+func TestSaveAndGetTranscript_RoundTrip(t *testing.T) {
+	s := newStore(t)
+	resetDB(t, rawPool(t))
+	ctx := context.Background()
+
+	want := domain.Transcript{Source: domain.SourceCaptions, Language: "en", Content: "the words"}
+	require.NoError(t, s.SaveTranscript(ctx, "youtube", "vid-1", want))
+
+	got, ok, err := s.GetTranscript(ctx, "youtube", "vid-1")
+	require.NoError(t, err)
+	require.True(t, ok, "a saved transcript must be found")
+	require.Equal(t, domain.SourceCaptions, got.Source)
+	require.Equal(t, "en", got.Language)
+	require.Equal(t, "the words", got.Content)
+	require.True(t, got.HasText())
+}
+
+func TestGetTranscript_Miss(t *testing.T) {
+	s := newStore(t)
+	resetDB(t, rawPool(t))
+
+	_, ok, err := s.GetTranscript(context.Background(), "youtube", "absent")
+	require.NoError(t, err, "a miss is not an error")
+	require.False(t, ok)
+}
+
+// A stored "no captions" outcome is a real hit: callers must skip without
+// re-fetching, so ok is true even though there is no text (ADR-021 / ADR-007).
+func TestSaveAndGetTranscript_NoneIsAStoredHit(t *testing.T) {
+	s := newStore(t)
+	resetDB(t, rawPool(t))
+	ctx := context.Background()
+
+	require.NoError(t, s.SaveTranscript(ctx, "youtube", "vid-none", domain.Transcript{Source: domain.SourceNone}))
+
+	got, ok, err := s.GetTranscript(ctx, "youtube", "vid-none")
+	require.NoError(t, err)
+	require.True(t, ok, "a stored SourceNone is a hit, not a miss")
+	require.Equal(t, domain.SourceNone, got.Source)
+	require.False(t, got.HasText())
+}
+
+// A transient 429 must never be persisted as a terminal transcript, or a later
+// read would mask the rate-limit as a permanent "no transcript" (ADR-014).
+func TestSaveTranscript_RejectsRateLimited(t *testing.T) {
+	s := newStore(t)
+	resetDB(t, rawPool(t))
+
+	err := s.SaveTranscript(context.Background(), "youtube", "vid-429",
+		domain.Transcript{Source: domain.SourceRateLimited})
+	require.Error(t, err)
+
+	_, ok, _ := s.GetTranscript(context.Background(), "youtube", "vid-429")
+	require.False(t, ok, "a rejected rate-limited save must leave nothing stored")
+}
+
+func TestSaveTranscript_UpsertLastWriteWins(t *testing.T) {
+	s := newStore(t)
+	resetDB(t, rawPool(t))
+	ctx := context.Background()
+
+	require.NoError(t, s.SaveTranscript(ctx, "youtube", "vid-up", domain.Transcript{Source: domain.SourceNone}))
+	require.NoError(t, s.SaveTranscript(ctx, "youtube", "vid-up",
+		domain.Transcript{Source: domain.SourceCaptions, Language: "en", Content: "now resolved"}))
+
+	got, ok, err := s.GetTranscript(ctx, "youtube", "vid-up")
+	require.NoError(t, err)
+	require.True(t, ok)
+	require.Equal(t, domain.SourceCaptions, got.Source)
+	require.Equal(t, "now resolved", got.Content)
+}

internal/adapters/store/videos.go

@@ -46,14 +46,15 @@ func (s *Store) UpsertVideo(ctx context.Context, v domain.Video) (string, error)
 		}
 
 		if err := tx.QueryRow(ctx,
-			`INSERT INTO videos (user_id, provider, provider_video_id, title, url, published_at)
+			`INSERT INTO videos (user_id, provider, provider_video_id, title, url, published_at, channel_title)
-			 VALUES ($1, $2, $3, $4, $5, $6)
+			 VALUES ($1, $2, $3, $4, $5, $6, $7)
 			 ON CONFLICT (user_id, provider, provider_video_id) DO UPDATE SET
 			     title         = EXCLUDED.title,
 			     url           = EXCLUDED.url,
-			     published_at = EXCLUDED.published_at
+			     published_at  = EXCLUDED.published_at,
+			     channel_title = COALESCE(NULLIF(EXCLUDED.channel_title, ''), videos.channel_title)
 			 RETURNING id`,
-			v.UserID, provider, v.ProviderVideoID, v.Title, v.URL, nullTime(v.PublishedAt),
+			v.UserID, provider, v.ProviderVideoID, v.Title, v.URL, nullTime(v.PublishedAt), v.ChannelTitle,
 		).Scan(&id); err != nil {
 			return fmt.Errorf("store: upsert video: %w", err)
 		}
@@ -72,3 +73,69 @@ func nullTime(t time.Time) *time.Time {
 	}
 	return &t
 }
+
+// NewestUnsummarizedVideoIDs returns up to limit of the user's videos that have
+// no summary yet, newest first (published_at DESC, NULLS LAST). It caps the
+// connect-time onboarding burst (Feature 1) at a fixed count: the caller marks
+// these for summarization through the shared rate gate. RLS-scoped via withUser,
+// so it only ever sees the requesting user's rows. limit <= 0 returns nil.
+func (s *Store) NewestUnsummarizedVideoIDs(ctx context.Context, userID string, limit int) ([]string, error) {
+	if limit <= 0 {
+		return nil, nil
+	}
+	var ids []string
+	if err := s.withUser(ctx, userID, func(tx pgx.Tx) error {
+		rows, err := tx.Query(ctx,
+			`SELECT v.id
+			   FROM videos v
+			  WHERE v.user_id = $1
+			    AND NOT EXISTS (
+			          SELECT 1 FROM summaries su
+			           WHERE su.user_id = v.user_id AND su.video_id = v.id)
+			  ORDER BY v.published_at DESC NULLS LAST, v.seen_at DESC
+			  LIMIT $2`, userID, limit)
+		if err != nil {
+			return fmt.Errorf("store: newest unsummarized: %w", err)
+		}
+		defer rows.Close()
+		for rows.Next() {
+			var id string
+			if err := rows.Scan(&id); err != nil {
+				return fmt.Errorf("store: scan newest unsummarized: %w", err)
+			}
+			ids = append(ids, id)
+		}
+		return rows.Err()
+	}); err != nil {
+		return nil, err
+	}
+	return ids, nil
+}
+
+// DistinctChannels returns the user's distinct, non-empty source channel titles
+// (the channels they have videos from), alphabetically — the option list for the
+// feed's channel filter. RLS-scoped via withUser.
+func (s *Store) DistinctChannels(ctx context.Context, userID string) ([]string, error) {
+	var out []string
+	if err := s.withUser(ctx, userID, func(tx pgx.Tx) error {
+		rows, err := tx.Query(ctx,
+			`SELECT DISTINCT channel_title FROM videos
+			  WHERE user_id = $1 AND channel_title IS NOT NULL AND channel_title <> ''
+			  ORDER BY channel_title`, userID)
+		if err != nil {
+			return fmt.Errorf("store: distinct channels: %w", err)
+		}
+		defer rows.Close()
+		for rows.Next() {
+			var c string
+			if err := rows.Scan(&c); err != nil {
+				return fmt.Errorf("store: scan channel: %w", err)
+			}
+			out = append(out, c)
+		}
+		return rows.Err()
+	}); err != nil {
+		return nil, err
+	}
+	return out, nil
+}

internal/adapters/store/videos_test.go

@@ -81,3 +81,61 @@ func TestUpsertVideo_PerUserIsolation(t *testing.T) {
 
 	require.NotEqual(t, idA, idB, "same provider video for two users must be two distinct rows")
 }
+
+func TestNewestUnsummarizedVideoIDs(t *testing.T) {
+	ctx := context.Background()
+	s := newStore(t)
+	resetDB(t, rawPool(t))
+
+	mk := func(user, pid string, day int) string {
+		v := ytVideo(user, pid, pid)
+		v.PublishedAt = time.Date(2026, 6, day, 12, 0, 0, 0, time.UTC)
+		id, err := s.UpsertVideo(ctx, v)
+		require.NoError(t, err)
+		return id
+	}
+
+	_ = mk(userA, "a1vid000001", 1)
+	id2 := mk(userA, "a2vid000002", 2)
+	id3 := mk(userA, "a3vid000003", 3)
+	id4 := mk(userA, "a4vid000004", 4)
+	mk(userB, "b1vid000009", 9) // userB's newest — must never leak via RLS
+
+	// The newest (v4) is summarized, so it's excluded from "unsummarized".
+	require.NoError(t, s.Deliver(ctx, summary(userA, id4, "done")))
+
+	// Cap 2, newest-first unsummarized: v3 then v2 (v4 excluded; userB excluded).
+	got, err := s.NewestUnsummarizedVideoIDs(ctx, userA, 2)
+	require.NoError(t, err)
+	require.Equal(t, []string{id3, id2}, got)
+
+	none, err := s.NewestUnsummarizedVideoIDs(ctx, userA, 0)
+	require.NoError(t, err)
+	require.Empty(t, none, "limit 0 returns nothing")
+}
+
+func TestUpsertVideoPersistsChannelAndDistinctChannels(t *testing.T) {
+	ctx := context.Background()
+	s := newStore(t)
+	resetDB(t, rawPool(t))
+
+	mk := func(pid, channel string) {
+		v := ytVideo(userA, pid, pid)
+		v.ChannelTitle = channel
+		_, err := s.UpsertVideo(ctx, v)
+		require.NoError(t, err)
+	}
+	mk("aa11111aaaa", "Acme Talks")
+	mk("bb22222bbbb", "Acme Talks") // same channel
+	mk("cc33333cccc", "Zeta Channel")
+	// userB's channel must not leak.
+	vb := ytVideo(userB, "dd44444dddd", "x")
+	vb.ChannelTitle = "Bravo Only"
+	_, err := s.UpsertVideo(ctx, vb)
+	require.NoError(t, err)
+
+	got, err := s.DistinctChannels(ctx, userA)
+	require.NoError(t, err)
+	require.Equal(t, []string{"Acme Talks", "Zeta Channel"}, got,
+		"distinct, alphabetical, user-scoped (no Bravo Only)")
+}

internal/adapters/summarizer/summarizer.go

@@ -1,17 +1,21 @@
 // Package summarizer implements ports.Summarizer backed by the copied llm
-// package's local-Primary -> BYO-Fallback routing (ADR-004). It is the only
+// package's routing (ADR-004, extended by ADR-022). It is the only place content
-// place content ever leaves the engine toward an AI model, so it is also the
+// ever leaves the engine toward an AI model, so it is also the enforcement point
-// enforcement point for the local-first guarantee in
+// for the local-first guarantee in docs/use-cases/ai_routing.feature: endpoints
-// docs/use-cases/ai_routing.feature: a user with no BYO provider configured has
+// are tried in order, locals first, so content only reaches an external model
-// their content sent to the local stack and nowhere else.
+// after every local endpoint has failed — and never at all when no external
+// endpoint is configured.
 package summarizer
 
 import (
+	"bytes"
 	"context"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"strings"
 	"time"
+	"unicode/utf8"
 
 	"gitea.d-ma.be/mathias/tapir/internal/domain"
 )
@@ -29,19 +33,41 @@ type Endpoint struct {
 	Model    string // resolved alias, e.g. "iguana/deepseek-r1-14b"
 }
 
-// Summarizer routes a transcript through the local endpoint first, then the
+// Summarizer routes a transcript through an ordered chain of endpoints, trying
-// optional BYO endpoint. It owns its routing (rather than delegating to
+// each in turn until one returns a parseable summary. It owns its routing
-// llm.Router) so it can record which provider answered and whether the fallback
+// (rather than delegating to llm.Router) so it can record which provider answered
-// was used — information llm.Router collapses away.
+// and whether a fallback was used — information llm.Router collapses away. The
+// chain ordering is the local-first guarantee: callers place local endpoints
+// first and any external endpoint last, so content only reaches an external model
+// after every local endpoint has failed.
 type Summarizer struct {
-	primary  Endpoint
+	endpoints     []Endpoint
-	fallback *Endpoint // nil => no BYO; primary errors are returned, never sent externally
+	maxInputChars int // transcript truncation budget; 0 = no limit
 	now           func() time.Time
 }
 
-// New constructs a Summarizer. fallback may be nil (no BYO provider configured).
+// New constructs a Summarizer from a primary endpoint and an optional fallback
+// (the historical local-Primary -> BYO-Fallback shape, ADR-004). A nil fallback
+// means a single-endpoint chain: errors are returned, content never leaves it.
 func New(primary Endpoint, fallback *Endpoint) *Summarizer {
-	return &Summarizer{primary: primary, fallback: fallback, now: time.Now}
+	eps := []Endpoint{primary}
+	if fallback != nil {
+		eps = append(eps, *fallback)
+	}
+	return &Summarizer{endpoints: eps, now: time.Now}
+}
+
+// NewChain constructs a Summarizer over an ordered endpoint chain (ADR-022).
+// endpoints are tried in order; the first to return a parseable summary wins, and
+// FallbackUsed is recorded true for any endpoint past the first. maxInputChars
+// bounds the transcript text sent to every endpoint (0 = unbounded), so a long
+// transcript does not overflow a small-context primary model's window. It panics
+// on an empty chain — a wiring bug, not a runtime condition.
+func NewChain(endpoints []Endpoint, maxInputChars int) *Summarizer {
+	if len(endpoints) == 0 {
+		panic("summarizer: NewChain requires at least one endpoint")
+	}
+	return &Summarizer{endpoints: endpoints, maxInputChars: maxInputChars, now: time.Now}
 }
 
 const systemPrompt = `You are Tapir, a video-summarization assistant.
@@ -53,31 +79,35 @@ Respond with ONLY a JSON object, no prose and no code fences:
 - "takeaways": the actionable conclusions a viewer should leave with.
 Output the JSON object and nothing else.`
 
-// Summarize implements ports.Summarizer.
+// Summarize implements ports.Summarizer. It walks the endpoint chain in order:
+// the first endpoint whose reply parses into a non-empty summary wins. An
+// endpoint is considered failed — and the next one tried — when the model call
+// errors OR when its reply cannot be parsed (a 200 with malformed JSON or a
+// highlights field the model emitted as a bare string). Truncation is applied
+// once, up front, so every endpoint sees the same bounded prompt. When the whole
+// chain fails, the joined error is returned so the engine queues the work for
+// retry and delivers no summary.
 func (s *Summarizer) Summarize(ctx context.Context, v domain.Video, t domain.Transcript) (domain.Summary, error) {
 	if !t.HasText() {
 		return domain.Summary{}, fmt.Errorf("summarize: transcript for video %s has no text", v.ID)
 	}
-	user := buildUserPrompt(v, t)
+	user := buildUserPrompt(v, t, s.maxInputChars)
 
-	// Primary = local stack. Only on its failure is anything sent externally,
+	var errs []error
-	// and only when a BYO fallback is configured.
+	for i, ep := range s.endpoints {
-	out, err := s.primary.Client.Complete(ctx, systemPrompt, user)
+		out, err := ep.Client.Complete(ctx, systemPrompt, user)
-	if err == nil {
+		if err != nil {
-		return s.build(v, s.primary, false, out)
+			errs = append(errs, fmt.Errorf("%s/%s call: %w", ep.Provider, ep.Model, err))
+			continue
 		}
-
+		sum, perr := s.build(v, ep, i > 0, out)
-	if s.fallback == nil {
+		if perr != nil {
-		// No BYO: content was sent to the local stack only. Surface the error so
+			errs = append(errs, fmt.Errorf("%s/%s output: %w", ep.Provider, ep.Model, perr))
-		// the engine can queue the work for retry; deliver no summary.
+			continue
-		return domain.Summary{}, fmt.Errorf("summarize: local AI failed and no BYO provider configured: %w", err)
 		}
-
+		return sum, nil
-	out, ferr := s.fallback.Client.Complete(ctx, systemPrompt, user)
-	if ferr != nil {
-		return domain.Summary{}, fmt.Errorf("summarize: local AI failed: %w; BYO %s failed: %v", err, s.fallback.Provider, ferr)
 	}
-	return s.build(v, *s.fallback, true, out)
+	return domain.Summary{}, fmt.Errorf("summarize: all %d endpoint(s) failed: %w", len(s.endpoints), errors.Join(errs...))
 }
 
 func (s *Summarizer) build(v domain.Video, ep Endpoint, fallbackUsed bool, raw string) (domain.Summary, error) {
@@ -89,8 +119,8 @@ func (s *Summarizer) build(v domain.Video, ep Endpoint, fallbackUsed bool, raw s
 		UserID:       v.UserID,
 		VideoID:      v.ID,
 		Summary:      parsed.Summary,
-		Highlights:   parsed.Highlights,
+		Highlights:   []string(parsed.Highlights),
-		Takeaways:    parsed.Takeaways,
+		Takeaways:    []string(parsed.Takeaways),
 		AIProvider:   ep.Provider,
 		AIModel:      ep.Model,
 		FallbackUsed: fallbackUsed,
@@ -98,20 +128,94 @@ func (s *Summarizer) build(v domain.Video, ep Endpoint, fallbackUsed bool, raw s
 	}, nil
 }
 
-func buildUserPrompt(v domain.Video, t domain.Transcript) string {
+func buildUserPrompt(v domain.Video, t domain.Transcript, maxInputChars int) string {
 	var b strings.Builder
 	fmt.Fprintf(&b, "Title: %s\n", v.Title)
 	if v.URL != "" {
 		fmt.Fprintf(&b, "URL: %s\n", v.URL)
 	}
-	fmt.Fprintf(&b, "\nTranscript:\n%s", t.Content)
+	fmt.Fprintf(&b, "\nTranscript:\n%s", truncate(t.Content, maxInputChars))
 	return b.String()
 }
 
+// truncate caps content to max bytes on a UTF-8 rune boundary, appending a
+// marker so the model knows the transcript was cut. A non-positive max (or a
+// content already within budget) returns content unchanged. Bounding the input
+// keeps a long transcript from overflowing a small-context model's window — the
+// production failure mode where koala/phi4-mini's 8k context returned HTTP 400 on
+// a 11.6k-token transcript.
+func truncate(content string, max int) string {
+	if max <= 0 || len(content) <= max {
+		return content
+	}
+	cut := max
+	for cut > 0 && !utf8.RuneStart(content[cut]) {
+		cut--
+	}
+	return content[:cut] + "\n…[transcript truncated to fit the model context]"
+}
+
+// flexStrings is a []string that also unmarshals from a single JSON string or a
+// JSON array of scalars. Small local models (koala/phi4-mini) sometimes emit
+// "highlights": "one point" instead of an array, or mix in a number; rather than
+// fail the whole summary on that quirk, coerce to []string. Empty/whitespace
+// elements are dropped.
+type flexStrings []string
+
+func (f *flexStrings) UnmarshalJSON(b []byte) error {
+	b = bytes.TrimSpace(b)
+	if len(b) == 0 || string(b) == "null" {
+		*f = nil
+		return nil
+	}
+	if b[0] == '[' {
+		var raw []json.RawMessage
+		if err := json.Unmarshal(b, &raw); err != nil {
+			return err
+		}
+		out := make([]string, 0, len(raw))
+		for _, r := range raw {
+			s, err := rawToString(r)
+			if err != nil {
+				return err
+			}
+			if strings.TrimSpace(s) != "" {
+				out = append(out, s)
+			}
+		}
+		*f = out
+		return nil
+	}
+	s, err := rawToString(b)
+	if err != nil {
+		return err
+	}
+	if strings.TrimSpace(s) == "" {
+		*f = nil
+	} else {
+		*f = flexStrings{s}
+	}
+	return nil
+}
+
+// rawToString renders a JSON scalar as text: a quoted string is unquoted; any
+// other scalar (number, bool) is kept as its literal source so no content is lost.
+func rawToString(r json.RawMessage) (string, error) {
+	r = bytes.TrimSpace(r)
+	if len(r) > 0 && r[0] == '"' {
+		var s string
+		if err := json.Unmarshal(r, &s); err != nil {
+			return "", err
+		}
+		return s, nil
+	}
+	return string(r), nil
+}
+
 type parsedSummary struct {
 	Summary    string      `json:"summary"`
-	Highlights []string `json:"highlights"`
+	Highlights flexStrings `json:"highlights"`
-	Takeaways  []string `json:"takeaways"`
+	Takeaways  flexStrings `json:"takeaways"`
 }
 
 // parse extracts the JSON object from a model reply. Thinking models (qwen3,

internal/adapters/summarizer/summarizer_test.go

@@ -129,8 +129,8 @@ func TestSummarize_NoBYO_ContentOnlyLocal(t *testing.T) {
 	local := &fakeClient{reply: goodReply}
 	s := New(Endpoint{Client: local, Provider: "local", Model: "iguana/deepseek-r1-14b"}, nil)
 
-	if s.fallback != nil {
+	if len(s.endpoints) != 1 {
-		t.Fatal("no BYO configured but fallback endpoint is non-nil")
+		t.Fatalf("no BYO configured but chain has %d endpoints, want 1", len(s.endpoints))
 	}
 	for i := 0; i < 3; i++ {
 		sum, err := s.Summarize(context.Background(), testVideo(), testTranscript())
@@ -176,3 +176,97 @@ func TestParse_EmptySummaryRejected(t *testing.T) {
 		t.Fatal("want error for empty summary (thinking model returned no content)")
 	}
 }
+
+// parse tolerates a small model emitting "highlights" as a bare string instead
+// of an array — the production koala/phi4-mini quirk that errored with
+// "cannot unmarshal string into Go struct field ... highlights of type []string".
+func TestParse_ToleratesStringHighlights(t *testing.T) {
+	p, err := parse(`{"summary":"s","highlights":"one big point","takeaways":["a","b"]}`)
+	if err != nil {
+		t.Fatalf("parse: %v", err)
+	}
+	if len(p.Highlights) != 1 || p.Highlights[0] != "one big point" {
+		t.Errorf("highlights = %v, want [\"one big point\"]", p.Highlights)
+	}
+	if len(p.Takeaways) != 2 {
+		t.Errorf("takeaways = %v, want 2", p.Takeaways)
+	}
+}
+
+// Chain: an endpoint that returns a 200 with unparseable output is treated as a
+// failure, and the next endpoint in the chain is tried. This is the case the old
+// primary->fallback shape missed — a parse error short-circuited instead of
+// falling back.
+func TestSummarize_FallsBackOnMalformedOutput(t *testing.T) {
+	bad := &fakeClient{reply: `{"summary": not json`}
+	good := &fakeClient{reply: goodReply}
+	s := NewChain([]Endpoint{
+		{Client: bad, Provider: "local", Model: "koala/phi4-mini"},
+		{Client: good, Provider: "local", Model: "koala/phi4-14b"},
+	}, 0)
+
+	sum, err := s.Summarize(context.Background(), testVideo(), testTranscript())
+	if err != nil {
+		t.Fatalf("Summarize: %v", err)
+	}
+	if sum.AIModel != "koala/phi4-14b" {
+		t.Errorf("AIModel = %q, want koala/phi4-14b (fell back past malformed primary)", sum.AIModel)
+	}
+	if !sum.FallbackUsed {
+		t.Error("FallbackUsed = false, want true")
+	}
+	if bad.calls != 1 || good.calls != 1 {
+		t.Errorf("calls: bad=%d good=%d, want 1 and 1", bad.calls, good.calls)
+	}
+}
+
+// Chain: when every endpoint fails, no summary is produced and the joined error
+// names each failure so the engine queues the work for retry.
+func TestSummarize_ChainAllEndpointsFail(t *testing.T) {
+	a := &fakeClient{err: errors.New("context overflow")}
+	b := &fakeClient{reply: "not even json"}
+	s := NewChain([]Endpoint{
+		{Client: a, Provider: "local", Model: "m1"},
+		{Client: b, Provider: "berget", Model: "m2"},
+	}, 0)
+
+	if _, err := s.Summarize(context.Background(), testVideo(), testTranscript()); err == nil {
+		t.Fatal("want error when all endpoints fail")
+	}
+	if a.calls != 1 || b.calls != 1 {
+		t.Errorf("calls: a=%d b=%d, want 1 and 1", a.calls, b.calls)
+	}
+}
+
+// A transcript longer than the chain's input budget is truncated before it
+// reaches any model, so a small-context primary does not overflow its window.
+func TestSummarize_TruncatesLongTranscript(t *testing.T) {
+	local := &fakeClient{reply: goodReply}
+	const budget = 100
+	s := NewChain([]Endpoint{{Client: local, Provider: "local", Model: "m"}}, budget)
+
+	long := domain.Transcript{
+		VideoID: "vid-1", UserID: "user-1", Source: domain.SourceCaptions,
+		Content: strings.Repeat("word ", 1000), // 5000 bytes, well over budget
+	}
+	if _, err := s.Summarize(context.Background(), testVideo(), long); err != nil {
+		t.Fatalf("Summarize: %v", err)
+	}
+	// The prompt carries title/URL framing plus the truncation marker, so allow
+	// headroom over the raw transcript budget — but it must be far below 5000.
+	if len(local.lastUser) > budget+300 {
+		t.Errorf("prompt length = %d, want <= %d (transcript not truncated)", len(local.lastUser), budget+300)
+	}
+	if !strings.Contains(local.lastUser, "truncated") {
+		t.Error("truncation marker missing from prompt")
+	}
+}
+
+func TestNewChain_PanicsOnEmptyChain(t *testing.T) {
+	defer func() {
+		if recover() == nil {
+			t.Fatal("want panic on empty endpoint chain")
+		}
+	}()
+	NewChain(nil, 0)
+}

internal/adapters/youtube/videobyid_test.go

@@ -0,0 +1,62 @@
+package youtube
+
+import (
+	"context"
+	"errors"
+	"net/http"
+	"testing"
+
+	"gitea.d-ma.be/mathias/tapir/internal/domain"
+)
+
+func TestVideoByID(t *testing.T) {
+	const id = "dQw4w9WgXcQ"
+	a, secrets := newTestAdapter(t, func(w http.ResponseWriter, r *http.Request) {
+		if r.URL.Path != "/videos" {
+			t.Errorf("unexpected path %q (must use videos.list)", r.URL.Path)
+		}
+		if got := r.URL.Query().Get("id"); got != id {
+			t.Errorf("expected id=%s, got %q", id, got)
+		}
+		if got := r.URL.Query().Get("part"); got != "snippet" {
+			t.Errorf("expected part=snippet, got %q", got)
+		}
+		_, _ = w.Write([]byte(`{"items":[{"snippet":{"title":"Never Gonna Give You Up","channelTitle":"Rick Astley","publishedAt":"2026-05-20T09:00:00Z"}}]}`))
+	})
+
+	v, err := a.VideoByID(context.Background(), "u1", id)
+	if err != nil {
+		t.Fatalf("VideoByID: %v", err)
+	}
+	if v.UserID != "u1" {
+		t.Errorf("UserID = %q, want u1", v.UserID)
+	}
+	if v.ProviderVideoID != id || v.Title != "Never Gonna Give You Up" {
+		t.Errorf("unexpected video: %+v", v)
+	}
+	if v.ChannelTitle != "Rick Astley" {
+		t.Errorf("ChannelTitle = %q, want Rick Astley", v.ChannelTitle)
+	}
+	if v.Provider != domain.ProviderYouTube || v.URL != "https://www.youtube.com/watch?v="+id {
+		t.Errorf("video not wired correctly: %+v", v)
+	}
+	if v.PublishedAt.IsZero() {
+		t.Errorf("expected publishedAt parsed, got zero")
+	}
+	if v.SubscriptionID != "" {
+		t.Errorf("a pasted video must have no subscription, got %q", v.SubscriptionID)
+	}
+	if secrets.byRef == nil {
+		t.Errorf("token must be resolved by reference through the SecretStore")
+	}
+}
+
+func TestVideoByIDNotFound(t *testing.T) {
+	a, _ := newTestAdapter(t, func(w http.ResponseWriter, _ *http.Request) {
+		_, _ = w.Write([]byte(`{"items":[]}`))
+	})
+	_, err := a.VideoByID(context.Background(), "u1", "missingvid0")
+	if !errors.Is(err, domain.ErrVideoNotFound) {
+		t.Fatalf("VideoByID for missing id = %v, want domain.ErrVideoNotFound", err)
+	}
+}

internal/adapters/youtube/youtube.go

@@ -234,6 +234,7 @@ func (a *Adapter) NewVideos(ctx context.Context, sub domain.Subscription) ([]dom
 			Provider:        domain.ProviderYouTube,
 			ProviderVideoID: vid,
 			Title:           item.Snippet.Title,
+			ChannelTitle:    sub.ChannelTitle,
 			URL:             "https://www.youtube.com/watch?v=" + vid,
 			PublishedAt:     item.Snippet.PublishedAt,
 		})
@@ -244,6 +245,38 @@ func (a *Adapter) NewVideos(ctx context.Context, sub domain.Subscription) ([]dom
 	return videos, nil
 }
 
+// VideoByID fetches a single video's metadata (videos.list, snippet) for an
+// arbitrary video id — including channels the user does not follow (paste-a-URL,
+// Feature 2). This is a Data API call (1 quota unit), NOT the rate-limited
+// caption path, so it is not gated: only the later transcript fetch goes through
+// globalFetchGate. UserID is set on the result and SubscriptionID is left empty
+// (a pasted video has no subscription parent). Returns ErrVideoNotFound when the
+// id resolves to no video.
+func (a *Adapter) VideoByID(ctx context.Context, userID, videoID string) (domain.Video, error) {
+	client, err := a.httpClient(ctx, a.cfg.TokenSecretRef)
+	if err != nil {
+		return domain.Video{}, err
+	}
+	q := url.Values{"part": {"snippet"}, "id": {videoID}}
+	var resp videoListResponse
+	if err := a.getJSON(ctx, client, "/videos", q, &resp); err != nil {
+		return domain.Video{}, fmt.Errorf("video by id %q: %w", videoID, err)
+	}
+	if len(resp.Items) == 0 {
+		return domain.Video{}, fmt.Errorf("video %q: %w", videoID, domain.ErrVideoNotFound)
+	}
+	it := resp.Items[0]
+	return domain.Video{
+		UserID:          userID,
+		Provider:        domain.ProviderYouTube,
+		ProviderVideoID: videoID,
+		Title:           it.Snippet.Title,
+		ChannelTitle:    it.Snippet.ChannelTitle,
+		URL:             "https://www.youtube.com/watch?v=" + videoID,
+		PublishedAt:     it.Snippet.PublishedAt,
+	}, nil
+}
+
 // uploadsPlaylistID derives a channel's uploads playlist id at zero API cost:
 // a standard channel id "UCxxxx" maps to uploads playlist "UUxxxx". Returns
 // ok=false for ids that don't follow this convention (caller falls back to
@@ -342,6 +375,16 @@ type playlistItemListResponse struct {
 	} `json:"items"`
 }
 
+type videoListResponse struct {
+	Items []struct {
+		Snippet struct {
+			Title        string    `json:"title"`
+			ChannelTitle string    `json:"channelTitle"`
+			PublishedAt  time.Time `json:"publishedAt"`
+		} `json:"snippet"`
+	} `json:"items"`
+}
+
 type channelListResponse struct {
 	Items []struct {
 		ContentDetails struct {

internal/adapters/youtube/youtube_test.go

@@ -131,7 +131,7 @@ func TestNewVideos(t *testing.T) {
 		}`))
 	})
 
-	sub := domain.Subscription{ID: "s1", UserID: "u1", ChannelID: "UC_acme"}
+	sub := domain.Subscription{ID: "s1", UserID: "u1", ChannelID: "UC_acme", ChannelTitle: "Acme Channel"}
 	vids, err := a.NewVideos(context.Background(), sub)
 	if err != nil {
 		t.Fatalf("NewVideos: %v", err)
@@ -143,6 +143,9 @@ func TestNewVideos(t *testing.T) {
 	if v.ProviderVideoID != "vid1" || v.Title != "Designing for Attention" {
 		t.Errorf("unexpected video: %+v", v)
 	}
+	if v.ChannelTitle != "Acme Channel" {
+		t.Errorf("ChannelTitle = %q, want Acme Channel", v.ChannelTitle)
+	}
 	if v.Provider != domain.ProviderYouTube || v.URL != "https://www.youtube.com/watch?v=vid1" {
 		t.Errorf("video not wired correctly: %+v", v)
 	}

internal/config/config.go

@@ -13,6 +13,7 @@ import (
 	"os"
 	"path/filepath"
 	"sort"
+	"strconv"
 	"strings"
 	"time"
 )
@@ -27,8 +28,24 @@ type Config struct {
 	GatewayURL string
 	// GatewayKey authorizes the gateway. Read from env, never committed.
 	GatewayKey string
-	// SummarizerModel is the alias in host/name form, e.g. "koala/phi4-mini".
+	// SummarizerModel is the primary summarizer alias in host/name form, tried
+	// first on every video, e.g. "koala/phi4-mini".
 	SummarizerModel string
+	// FallbackModel is the LOCAL fallback alias tried when the primary fails or
+	// returns unparseable output (ADR-022). Kept local so content stays on the
+	// homelab stack. Empty disables it. Default a bigger-context local model.
+	FallbackModel string
+	// CloudFallbackModel is the worst-case EXTERNAL fallback alias, tried only
+	// after every local endpoint has failed (ADR-022). For client deployments set
+	// this empty so content never leaves the local stack. Default a berget alias.
+	CloudFallbackModel string
+	// SummaryMaxTokens caps the completion budget per summary call. Small-context
+	// models (koala/phi4-mini, 8k) overflow when prompt + max_tokens exceeds the
+	// window; a summary needs only a few hundred tokens, so the default is small.
+	SummaryMaxTokens int
+	// MaxTranscriptChars bounds the transcript text sent to the model so a long
+	// transcript does not overflow a small-context primary. 0 disables truncation.
+	MaxTranscriptChars int
 	// SummarizerTimeout bounds a single completion call. Thinking models are
 	// slow, so the default is generous.
 	SummarizerTimeout time.Duration
@@ -79,6 +96,13 @@ type Config struct {
 	// pre-recency behaviour). Default ~7 days.
 	AutoSummarizeWindow time.Duration
 
+	// OnboardSummarizeCount caps how many of a freshly-connected user's newest
+	// videos are summarized immediately on connect (the onboarding "it works"
+	// burst). HARD-capped at maxOnboardSummarizeCount so onboarding can never
+	// bulk-fetch; 0 disables the burst. Every fetch still flows through the shared
+	// caption rate gate (ADR-014) — the cap bounds count, never the pacing. Default 3.
+	OnboardSummarizeCount int
+
 	// DiscoveryInterval, when > 0, makes `serve` run in-process scheduled discovery
 	// for ALL users on that cadence (ADR-018). Zero/unset = disabled, so dev and
 	// tests never auto-fetch. Single-replica assumption — see cmdServe.
@@ -110,6 +134,10 @@ func (c Config) DexConfigured() bool { return strings.TrimSpace(c.OIDCIssuer) !=
 const (
 	defaultGatewayURL            = "http://koala:30401/v1"
 	defaultSummarizerModel       = "koala/phi4-mini"
+	defaultFallbackModel         = "koala/phi4-14b"
+	defaultCloudFallbackModel    = "berget/mistral-small"
+	defaultSummaryMaxTokens      = 1500
+	defaultMaxTranscriptChars    = 18000
 	defaultSummarizerTimeout     = 5 * time.Minute
 	defaultYTTokenRef            = "youtube/refresh_token"
 	defaultYTConnectRedirectURL  = "https://tapir.d-ma.be/oauth/youtube/callback"
@@ -119,6 +147,8 @@ const (
 	defaultFetchRate             = 2 * time.Second
 	defaultPublicURL             = "https://tapir.d-ma.be"
 	defaultAutoSummarizeWindow   = 7 * 24 * time.Hour
+	defaultOnboardSummarizeCount = 3
+	maxOnboardSummarizeCount     = 5
 )
 
 // Load reads the environment into a Config, applying defaults. It does not
@@ -131,6 +161,8 @@ func Load() (Config, error) {
 		GatewayURL:           envOr("TAPIR_GATEWAY_URL", defaultGatewayURL),
 		GatewayKey:           os.Getenv("TAPIR_GATEWAY_KEY"),
 		SummarizerModel:      envOr("TAPIR_SUMMARIZER_MODEL", defaultSummarizerModel),
+		FallbackModel:        lookupOr("TAPIR_FALLBACK_MODEL", defaultFallbackModel),
+		CloudFallbackModel:   lookupOr("TAPIR_CLOUD_FALLBACK_MODEL", defaultCloudFallbackModel),
 		DBDSN:                os.Getenv("TAPIR_DB_DSN"),
 		YTClientID:           os.Getenv("TAPIR_YT_CLIENT_ID"),
 		YTClientSecret:       os.Getenv("TAPIR_YT_CLIENT_SECRET"),
@@ -183,6 +215,33 @@ func Load() (Config, error) {
 	}
 	c.AutoSummarizeWindow = autoWindow
 
+	summaryTokens, err := intOr("TAPIR_SUMMARY_MAX_TOKENS", defaultSummaryMaxTokens)
+	if err != nil {
+		return Config{}, err
+	}
+	c.SummaryMaxTokens = summaryTokens
+
+	maxChars, err := intOr("TAPIR_MAX_TRANSCRIPT_CHARS", defaultMaxTranscriptChars)
+	if err != nil {
+		return Config{}, err
+	}
+	if maxChars < 0 {
+		maxChars = 0
+	}
+	c.MaxTranscriptChars = maxChars
+
+	onboard, err := intOr("TAPIR_ONBOARD_SUMMARIZE_COUNT", defaultOnboardSummarizeCount)
+	if err != nil {
+		return Config{}, err
+	}
+	if onboard < 0 {
+		onboard = 0
+	}
+	if onboard > maxOnboardSummarizeCount {
+		onboard = maxOnboardSummarizeCount
+	}
+	c.OnboardSummarizeCount = onboard
+
 	return c, nil
 }
 
@@ -247,6 +306,29 @@ func envOr(key, fallback string) string {
 	return fallback
 }
 
+// lookupOr returns the env value when the key is PRESENT (even if empty), else
+// fallback. Unlike envOr it lets an explicit empty value override the default —
+// needed to DISABLE an optional fallback model (e.g. set the cloud fallback empty
+// for a client deployment so content never leaves the local stack).
+func lookupOr(key, fallback string) string {
+	if v, ok := os.LookupEnv(key); ok {
+		return v
+	}
+	return fallback
+}
+
+func intOr(key string, fallback int) (int, error) {
+	v := os.Getenv(key)
+	if v == "" {
+		return fallback, nil
+	}
+	n, err := strconv.Atoi(v)
+	if err != nil {
+		return 0, fmt.Errorf("config: %s=%q: %w", key, v, err)
+	}
+	return n, nil
+}
+
 func durationOr(key string, fallback time.Duration) (time.Duration, error) {
 	v := os.Getenv(key)
 	if v == "" {

internal/config/config_test.go

@@ -1,11 +1,29 @@
 package config
 
 import (
+	"os"
 	"strings"
 	"testing"
 	"time"
 )
 
+// unset removes an env key for the duration of the test, restoring it after.
+// Needed to observe a default for a key read with LookupEnv (where present-empty
+// means "explicitly disabled", not "use default").
+func unset(t *testing.T, key string) {
+	t.Helper()
+	if old, ok := os.LookupEnv(key); ok {
+		t.Cleanup(func() {
+			if err := os.Setenv(key, old); err != nil {
+				t.Fatalf("restore %s: %v", key, err)
+			}
+		})
+	}
+	if err := os.Unsetenv(key); err != nil {
+		t.Fatalf("unset %s: %v", key, err)
+	}
+}
+
 // setEnv sets env vars for the test and clears them afterward, so cases don't
 // leak into one another. t.Setenv handles restoration.
 func setEnv(t *testing.T, kv map[string]string) {
@@ -53,6 +71,78 @@ func TestLoad_AppliesDefaults(t *testing.T) {
 	}
 }
 
+func TestLoad_SummarizerChainDefaults(t *testing.T) {
+	setEnv(t, map[string]string{
+		"TAPIR_SUMMARIZER_MODEL":     "",
+		"TAPIR_SUMMARY_MAX_TOKENS":   "",
+		"TAPIR_MAX_TRANSCRIPT_CHARS": "",
+	})
+	unset(t, "TAPIR_FALLBACK_MODEL")
+	unset(t, "TAPIR_CLOUD_FALLBACK_MODEL")
+
+	c, err := Load()
+	if err != nil {
+		t.Fatalf("Load: %v", err)
+	}
+	if c.FallbackModel != defaultFallbackModel {
+		t.Errorf("FallbackModel = %q, want %q", c.FallbackModel, defaultFallbackModel)
+	}
+	if c.CloudFallbackModel != defaultCloudFallbackModel {
+		t.Errorf("CloudFallbackModel = %q, want %q", c.CloudFallbackModel, defaultCloudFallbackModel)
+	}
+	if c.SummaryMaxTokens != defaultSummaryMaxTokens {
+		t.Errorf("SummaryMaxTokens = %d, want %d", c.SummaryMaxTokens, defaultSummaryMaxTokens)
+	}
+	if c.MaxTranscriptChars != defaultMaxTranscriptChars {
+		t.Errorf("MaxTranscriptChars = %d, want %d", c.MaxTranscriptChars, defaultMaxTranscriptChars)
+	}
+}
+
+// An explicitly empty cloud-fallback env disables external routing — the lever a
+// client deployment pulls so content never leaves the local stack.
+func TestLoad_EmptyCloudFallbackDisables(t *testing.T) {
+	t.Setenv("TAPIR_CLOUD_FALLBACK_MODEL", "")
+	c, err := Load()
+	if err != nil {
+		t.Fatalf("Load: %v", err)
+	}
+	if c.CloudFallbackModel != "" {
+		t.Errorf("CloudFallbackModel = %q, want empty (disabled)", c.CloudFallbackModel)
+	}
+}
+
+func TestLoad_OnboardSummarizeCount(t *testing.T) {
+	cases := []struct {
+		name, env string
+		want      int
+	}{
+		{"default", "", defaultOnboardSummarizeCount},
+		{"explicit", "4", 4},
+		{"zero disables", "0", 0},
+		{"clamped to hard cap", "50", maxOnboardSummarizeCount},
+		{"negative clamps to zero", "-3", 0},
+	}
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			setEnv(t, map[string]string{"TAPIR_ONBOARD_SUMMARIZE_COUNT": c.env})
+			cfg, err := Load()
+			if err != nil {
+				t.Fatalf("Load: %v", err)
+			}
+			if cfg.OnboardSummarizeCount != c.want {
+				t.Fatalf("OnboardSummarizeCount = %d, want %d", cfg.OnboardSummarizeCount, c.want)
+			}
+		})
+	}
+}
+
+func TestLoad_OnboardSummarizeCountInvalid(t *testing.T) {
+	setEnv(t, map[string]string{"TAPIR_ONBOARD_SUMMARIZE_COUNT": "three"})
+	if _, err := Load(); err == nil {
+		t.Fatal("Load: want error for non-numeric TAPIR_ONBOARD_SUMMARIZE_COUNT")
+	}
+}
+
 func TestLoad_ParsesValues(t *testing.T) {
 	setEnv(t, map[string]string{
 		"TAPIR_USER_ID":               "11111111-1111-1111-1111-111111111111",

internal/domain/domain.go

@@ -3,10 +3,16 @@
 package domain
 
 import (
+	"errors"
 	"fmt"
 	"time"
 )
 
+// ErrVideoNotFound is returned when a video id resolves to no video (deleted,
+// private, or a typo'd paste). Defined in domain so adapters and the web layer
+// share one sentinel without coupling to each other.
+var ErrVideoNotFound = errors.New("video not found")
+
 // ErrChannelUnavailable is returned by a VideoSource when a channel's upload
 // playlist returns HTTP 404 — the channel was deleted or made private. The runner
 // stores these so the account page can surface them to the user.
@@ -67,6 +73,7 @@ type Video struct {
 	Provider        Provider
 	ProviderVideoID string
 	Title           string
+	ChannelTitle    string
 	URL             string
 	PublishedAt     time.Time
 	SeenAt          time.Time

internal/ports/ports.go

@@ -27,6 +27,26 @@ type Summarizer interface {
 	Summarize(ctx context.Context, v domain.Video, t domain.Transcript) (domain.Summary, error)
 }
 
+// TranscriptStore persists transcripts as shared, video-keyed public content
+// (ADR-021). It is keyed by the cross-user dedup key (provider, providerVideoID)
+// — the video's public identity, NOT Tapir's per-user videos.id — and holds only
+// public caption content, so it is deliberately NOT user-scoped: two users who
+// share a video share the one row. The engine reads it before any caption fetch
+// so re-analysis never re-touches YouTube (ADR-010/014).
+type TranscriptStore interface {
+	// GetTranscript returns the stored transcript for a video and whether one
+	// exists. A stored Source == SourceNone (captions permanently absent) is a
+	// real hit: ok is true and HasText() is false, so callers skip without
+	// re-fetching. A transient rate-limit is never stored, so it never appears
+	// here as a false absence.
+	GetTranscript(ctx context.Context, provider, providerVideoID string) (t domain.Transcript, ok bool, err error)
+	// SaveTranscript upserts the transcript for (provider, providerVideoID). Only
+	// terminal outcomes are persisted: SourceCaptions (with text) or SourceNone.
+	// SourceRateLimited must NOT be passed — it is a per-user retry (ADR-014), not
+	// a shared terminal state.
+	SaveTranscript(ctx context.Context, provider, providerVideoID string, t domain.Transcript) error
+}
+
 // Sink delivers a summary to a destination (user store, brain, ...).
 // Implementations fail independently of one another.
 type Sink interface {

internal/usecase/engine.go

@@ -27,6 +27,13 @@ type Engine struct {
 	AI     ports.Summarizer
 	Sinks  []ports.Sink
 
+	// Transcripts, when set, is the shared transcript cache (ADR-021): the engine
+	// reads it before any caption fetch and writes resolved transcripts back, so
+	// re-analysis — the same user re-summarizing, or a second user with the same
+	// video — never re-touches YouTube (ADR-010/014). Optional: nil disables
+	// persistence (fetch every time), keeping the pure-core/scaffold wiring valid.
+	Transcripts ports.TranscriptStore
+
 	// processed dedups videos within this engine's lifetime so a video is not
 	// summarized twice when the watcher sees it again. Durable cross-restart
 	// dedup is the store's concern (a resolved TRANSCRIPT / existing SUMMARY,
@@ -57,9 +64,9 @@ type ProcessResult struct {
 // resolve transcript -> (summarize -> deliver) | skip.
 // See docs/use-cases/summarize_new_video.feature.
 func (e *Engine) ProcessNewVideo(ctx context.Context, v domain.Video) (ProcessResult, error) {
-	t, err := e.Source.FetchTranscript(ctx, v)
+	t, err := e.resolveTranscript(ctx, v)
 	if err != nil {
-		return ProcessResult{Video: v}, fmt.Errorf("fetch transcript: %w", err)
+		return ProcessResult{Video: v}, err
 	}
 	if !t.HasText() {
 		// No usable transcript: record the skip, produce no summary, deliver nothing
@@ -86,6 +93,39 @@ func (e *Engine) ProcessNewVideo(ctx context.Context, v domain.Video) (ProcessRe
 	return ProcessResult{Video: v, Summary: &sum, TranscriptSource: string(t.Source)}, errors.Join(errs...)
 }
 
+// resolveTranscript returns v's transcript, reading the shared store first
+// (ADR-021): a stored transcript — including a stored SourceNone (captions
+// permanently absent) — is returned without touching YouTube, so re-analysis
+// never re-fetches. On a store miss it fetches through the source (which gates
+// the caption call, ADR-014) and persists the terminal outcome so the next
+// analysis, for any user, reads from the store. A transient SourceRateLimited is
+// returned to the caller (the runner stamps a per-user backoff) but never stored,
+// so persistence can never mask a 429 as a permanent "no transcript". When no
+// TranscriptStore is wired the engine simply fetches every time.
+func (e *Engine) resolveTranscript(ctx context.Context, v domain.Video) (domain.Transcript, error) {
+	if e.Transcripts != nil {
+		stored, ok, err := e.Transcripts.GetTranscript(ctx, string(v.Provider), v.ProviderVideoID)
+		if err != nil {
+			return domain.Transcript{}, fmt.Errorf("get stored transcript: %w", err)
+		}
+		if ok {
+			return stored, nil
+		}
+	}
+
+	t, err := e.Source.FetchTranscript(ctx, v)
+	if err != nil {
+		return domain.Transcript{}, fmt.Errorf("fetch transcript: %w", err)
+	}
+
+	if e.Transcripts != nil && t.Source != domain.SourceRateLimited {
+		if err := e.Transcripts.SaveTranscript(ctx, string(v.Provider), v.ProviderVideoID, t); err != nil {
+			return domain.Transcript{}, fmt.Errorf("save transcript: %w", err)
+		}
+	}
+	return t, nil
+}
+
 // ProcessNewVideos walks a user's subscriptions and processes each newly seen
 // video. Only videos surfaced via the user's subscriptions are considered, so a
 // channel the user is not subscribed to is never processed. A video already

internal/usecase/engine_transcript_test.go

@@ -0,0 +1,187 @@
+package usecase
+
+import (
+	"context"
+	"testing"
+
+	"gitea.d-ma.be/mathias/tapir/internal/domain"
+)
+
+// These tests pin the ADR-021 read-stored-first behaviour at the engine core:
+// a stored transcript is summarized without re-touching the source, a miss
+// fetches once and persists, and a transient rate-limit is never cached.
+
+type recordingSource struct {
+	transcript domain.Transcript
+	fetchCalls int
+}
+
+func (s *recordingSource) ListSubscriptions(context.Context, string) ([]domain.Subscription, error) {
+	return nil, nil
+}
+
+func (s *recordingSource) NewVideos(context.Context, domain.Subscription) ([]domain.Video, error) {
+	return nil, nil
+}
+
+func (s *recordingSource) FetchTranscript(context.Context, domain.Video) (domain.Transcript, error) {
+	s.fetchCalls++
+	return s.transcript, nil
+}
+
+type fakeTranscriptStore struct {
+	stored map[string]domain.Transcript
+	saves  int
+}
+
+func newFakeTranscriptStore() *fakeTranscriptStore {
+	return &fakeTranscriptStore{stored: make(map[string]domain.Transcript)}
+}
+
+func (f *fakeTranscriptStore) key(provider, id string) string { return provider + "|" + id }
+
+func (f *fakeTranscriptStore) GetTranscript(_ context.Context, provider, id string) (domain.Transcript, bool, error) {
+	t, ok := f.stored[f.key(provider, id)]
+	return t, ok, nil
+}
+
+func (f *fakeTranscriptStore) SaveTranscript(_ context.Context, provider, id string, t domain.Transcript) error {
+	f.saves++
+	f.stored[f.key(provider, id)] = t
+	return nil
+}
+
+type countingSummarizer struct{ calls int }
+
+func (c *countingSummarizer) Summarize(_ context.Context, v domain.Video, _ domain.Transcript) (domain.Summary, error) {
+	c.calls++
+	return domain.Summary{VideoID: v.ID, UserID: v.UserID, Summary: "s", AIProvider: "local"}, nil
+}
+
+type nopSink struct{}
+
+func (nopSink) Name() string                                  { return "nop" }
+func (nopSink) Deliver(context.Context, domain.Summary) error { return nil }
+
+func testVideo() domain.Video {
+	return domain.Video{ID: "v1", UserID: "u1", Provider: domain.ProviderYouTube, ProviderVideoID: "yt1"}
+}
+
+func TestProcessNewVideo_StoredTranscriptSkipsFetch(t *testing.T) {
+	src := &recordingSource{}
+	ts := newFakeTranscriptStore()
+	ts.stored[ts.key("youtube", "yt1")] = domain.Transcript{Source: domain.SourceCaptions, Content: "stored words"}
+	sum := &countingSummarizer{}
+	eng := NewEngine(src, sum, nopSink{})
+	eng.Transcripts = ts
+
+	res, err := eng.ProcessNewVideo(context.Background(), testVideo())
+	if err != nil {
+		t.Fatalf("ProcessNewVideo: %v", err)
+	}
+	if src.fetchCalls != 0 {
+		t.Fatalf("stored transcript must not re-fetch from source; got %d fetches", src.fetchCalls)
+	}
+	if ts.saves != 0 {
+		t.Fatalf("a store hit must not re-save; got %d saves", ts.saves)
+	}
+	if sum.calls != 1 || res.Summary == nil {
+		t.Fatalf("expected a summary from the stored transcript; calls=%d summary=%v", sum.calls, res.Summary)
+	}
+}
+
+func TestProcessNewVideo_StoreMissFetchesAndPersists(t *testing.T) {
+	src := &recordingSource{transcript: domain.Transcript{Source: domain.SourceCaptions, Language: "en", Content: "fetched words"}}
+	ts := newFakeTranscriptStore()
+	sum := &countingSummarizer{}
+	eng := NewEngine(src, sum, nopSink{})
+	eng.Transcripts = ts
+
+	if _, err := eng.ProcessNewVideo(context.Background(), testVideo()); err != nil {
+		t.Fatalf("ProcessNewVideo: %v", err)
+	}
+	if src.fetchCalls != 1 {
+		t.Fatalf("a store miss must fetch exactly once; got %d", src.fetchCalls)
+	}
+	if ts.saves != 1 {
+		t.Fatalf("a fetched transcript must be persisted; got %d saves", ts.saves)
+	}
+	got, ok, _ := ts.GetTranscript(context.Background(), "youtube", "yt1")
+	if !ok || got.Content != "fetched words" {
+		t.Fatalf("persisted transcript not readable back: ok=%v content=%q", ok, got.Content)
+	}
+}
+
+// The second summarize of the same video reads the persisted transcript and does
+// NOT re-fetch — the primary ADR-021 win, proven end to end at the engine.
+func TestProcessNewVideo_SecondSummarizeDoesNotRefetch(t *testing.T) {
+	src := &recordingSource{transcript: domain.Transcript{Source: domain.SourceCaptions, Content: "words"}}
+	ts := newFakeTranscriptStore()
+	eng := NewEngine(src, &countingSummarizer{}, nopSink{})
+	eng.Transcripts = ts
+
+	for i := 0; i < 2; i++ {
+		if _, err := eng.ProcessNewVideo(context.Background(), testVideo()); err != nil {
+			t.Fatalf("pass %d: %v", i, err)
+		}
+	}
+	if src.fetchCalls != 1 {
+		t.Fatalf("the second summarize must reuse the stored transcript; got %d fetches", src.fetchCalls)
+	}
+}
+
+// A stored "no captions" outcome short-circuits before both fetch and summarize.
+func TestProcessNewVideo_StoredNoneSkipsFetchAndSummarize(t *testing.T) {
+	src := &recordingSource{}
+	ts := newFakeTranscriptStore()
+	ts.stored[ts.key("youtube", "yt1")] = domain.Transcript{Source: domain.SourceNone}
+	sum := &countingSummarizer{}
+	eng := NewEngine(src, sum, nopSink{})
+	eng.Transcripts = ts
+
+	res, err := eng.ProcessNewVideo(context.Background(), testVideo())
+	if err != nil {
+		t.Fatalf("ProcessNewVideo: %v", err)
+	}
+	if !res.Skipped {
+		t.Fatal("a stored SourceNone must skip")
+	}
+	if src.fetchCalls != 0 || sum.calls != 0 {
+		t.Fatalf("stored none must neither fetch nor summarize; fetches=%d calls=%d", src.fetchCalls, sum.calls)
+	}
+}
+
+// A transient 429 is surfaced (so the runner backs off per-user) but never cached
+// as a shared terminal state — otherwise it would mask a rate-limit as permanent.
+func TestProcessNewVideo_RateLimitedIsNotPersisted(t *testing.T) {
+	src := &recordingSource{transcript: domain.Transcript{Source: domain.SourceRateLimited}}
+	ts := newFakeTranscriptStore()
+	eng := NewEngine(src, &countingSummarizer{}, nopSink{})
+	eng.Transcripts = ts
+
+	res, err := eng.ProcessNewVideo(context.Background(), testVideo())
+	if err != nil {
+		t.Fatalf("ProcessNewVideo: %v", err)
+	}
+	if !res.Skipped || res.TranscriptSource != string(domain.SourceRateLimited) {
+		t.Fatalf("expected a rate-limited skip; skipped=%v source=%q", res.Skipped, res.TranscriptSource)
+	}
+	if ts.saves != 0 {
+		t.Fatalf("a transient rate-limit must not be persisted; got %d saves", ts.saves)
+	}
+}
+
+// With no TranscriptStore wired the engine fetches every time (back-compat).
+func TestProcessNewVideo_NilStoreFetchesEveryTime(t *testing.T) {
+	src := &recordingSource{transcript: domain.Transcript{Source: domain.SourceCaptions, Content: "words"}}
+	eng := NewEngine(src, &countingSummarizer{}, nopSink{})
+
+	for i := 0; i < 2; i++ {
+		if _, err := eng.ProcessNewVideo(context.Background(), testVideo()); err != nil {
+			t.Fatalf("pass %d: %v", i, err)
+		}
+	}
+	if src.fetchCalls != 2 {
+		t.Fatalf("nil store must fetch every time; got %d", src.fetchCalls)
+	}
+}

internal/web/connect.go

@@ -23,6 +23,15 @@ type Connections interface {
 	UpsertConnection(ctx context.Context, userID string, c store.Connection) error
 }
 
+// DiscoveryTrigger requests an out-of-band discovery pass for a user. The connect
+// flow fires it the moment a YouTube account is linked so videos appear promptly
+// instead of waiting for the next scheduled pass (#6). Enqueue must be
+// non-blocking and safe to call from the request goroutine; the implementation
+// owns serialization with the scheduler (one pass at a time). nil = no trigger.
+type DiscoveryTrigger interface {
+	Enqueue(userID string)
+}
+
 // connectStateTTL bounds how long a generated CSRF state is valid between the
 // connect redirect and the provider callback.
 const connectStateTTL = 10 * time.Minute
@@ -43,6 +52,10 @@ type ConnectHandler struct {
 	Conns   Connections
 	Log     *slog.Logger
 
+	// Discovery, when set, is fired after a successful connect so the new
+	// connection's videos are discovered immediately (#6). Optional.
+	Discovery DiscoveryTrigger
+
 	states *connectStateStore
 	now    func() time.Time
 }
@@ -134,6 +147,12 @@ func (h *ConnectHandler) handleCallback(w http.ResponseWriter, r *http.Request)
 		return
 	}
 
+	// Discover this user's videos now rather than waiting for the next scheduled
+	// pass (#6). Non-blocking; the trigger serializes with the scheduler.
+	if h.Discovery != nil {
+		h.Discovery.Enqueue(userID)
+	}
+
 	setFlash(w, flashConnected)
 	http.Redirect(w, r, "/", http.StatusSeeOther)
 }

internal/web/connect_test.go

@@ -45,6 +45,24 @@ func (c *fakeConns) UpsertConnection(_ context.Context, userID string, conn stor
 	return nil
 }
 
+// fakeTrigger records Enqueue calls so a test can assert connect fired discovery.
+type fakeTrigger struct {
+	mu    sync.Mutex
+	users []string
+}
+
+func (f *fakeTrigger) Enqueue(userID string) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	f.users = append(f.users, userID)
+}
+
+func (f *fakeTrigger) seen() []string {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	return append([]string(nil), f.users...)
+}
+
 // tokenServer fakes Google's token endpoint, returning body for any POST.
 func tokenServer(t *testing.T, body string) *httptest.Server {
 	t.Helper()
@@ -126,6 +144,36 @@ func TestCallbackExchangesAndRecordsConnection(t *testing.T) {
 	require.Equal(t, wantRef, conns.conn.TokenRef)
 }
 
+func TestCallbackTriggersDiscovery(t *testing.T) {
+	srv := tokenServer(t,
+		`{"access_token":"at","refresh_token":"rt-secret","token_type":"Bearer","expires_in":3600}`)
+	app := newConnectApp(t, srv.URL, &fakeWriter{}, &fakeConns{})
+	trig := &fakeTrigger{}
+	app.Connect.Discovery = trig
+
+	state := connectState(t, app)
+	rec := do(t, app, httptest.NewRequest(http.MethodGet,
+		"/oauth/youtube/callback?state="+state+"&code=the-code", nil))
+
+	require.Equal(t, http.StatusSeeOther, rec.Code)
+	require.Equal(t, []string{userID}, trig.seen(),
+		"a successful connect must trigger discovery for the connecting user")
+}
+
+func TestCallbackNoDiscoveryOnFailedConnect(t *testing.T) {
+	srv := tokenServer(t,
+		`{"access_token":"at","refresh_token":"rt","token_type":"Bearer","expires_in":3600}`)
+	app := newConnectApp(t, srv.URL, &fakeWriter{}, &fakeConns{})
+	trig := &fakeTrigger{}
+	app.Connect.Discovery = trig
+
+	// No state → CSRF reject → nothing connected, so no discovery.
+	rec := do(t, app, httptest.NewRequest(http.MethodGet,
+		"/oauth/youtube/callback?code=the-code", nil))
+	require.Equal(t, http.StatusBadRequest, rec.Code)
+	require.Empty(t, trig.seen(), "a failed connect must not trigger discovery")
+}
+
 func TestCallbackRejectsMissingState(t *testing.T) {
 	srv := tokenServer(t,
 		`{"access_token":"at","refresh_token":"rt","token_type":"Bearer","expires_in":3600}`)

internal/web/handlers.go

@@ -3,6 +3,8 @@ package web
 import (
 	"context"
 	"errors"
+	"html/template"
+	"io"
 	"log/slog"
 	"net/http"
 	"time"
@@ -10,6 +12,7 @@ import (
 	"github.com/a-h/templ"
 
 	"gitea.d-ma.be/mathias/tapir/internal/adapters/store"
+	"gitea.d-ma.be/mathias/tapir/internal/domain"
 )
 
 // Store is the read/write surface the web handlers depend on — a narrow port over
@@ -18,6 +21,9 @@ import (
 // fake without a database.
 type Store interface {
 	ListVideos(ctx context.Context, userID string, limit int) ([]store.SummaryRow, error)
+	// DistinctChannels lists the user's source channels — the options for the
+	// feed's channel multi-select filter.
+	DistinctChannels(ctx context.Context, userID string) ([]string, error)
 	GetSummaryByVideo(ctx context.Context, userID, videoID string) (*store.SummaryRow, error)
 	GetVideoRow(ctx context.Context, userID, videoID string) (*store.SummaryRow, error)
 	ActionsFor(ctx context.Context, userID string, videoIDs []string) (map[string][]string, error)
@@ -30,6 +36,10 @@ type Store interface {
 	SetAutoSummarize(ctx context.Context, userID string, enabled bool) error
 	RequestSummarize(ctx context.Context, userID, videoID string) error
 
+	// UpsertVideo persists a pasted video (idempotent on user+provider+video id,
+	// so it also dedups) and returns its durable store id.
+	UpsertVideo(ctx context.Context, v domain.Video) (string, error)
+
 	// Account management (the /account page, disconnect, delete-account).
 	ConnectionsForUser(ctx context.Context, userID string) ([]store.Connection, error)
 	DeleteConnection(ctx context.Context, userID, provider string) error
@@ -78,6 +88,9 @@ type App struct {
 	// background goroutine (the "Summarize" button kicks it off). Nil = queue-only:
 	// the button flips the DB flag and the next `tapir run` does the work.
 	Processor Processor
+	// Fetcher, when non-nil, resolves an arbitrary YouTube video id to metadata for
+	// the paste-a-URL flow (Feature 2). Nil = the /paste route is not mounted.
+	Fetcher VideoFetcher
 	// Processing tracks in-flight immediate summarizations so the status endpoint
 	// shows the animation until the summary lands. The zero value is ready to use.
 	Processing ProcessingSet
@@ -132,6 +145,9 @@ func (a *App) Router() http.Handler {
 	app.HandleFunc("POST /v/{videoId}/action", a.handleAction)
 	app.HandleFunc("POST /v/{videoId}/summarize", a.handleRequestSummarize)
 	app.HandleFunc("POST /v/{videoId}/retry-now", a.handleRetryNow)
+	if a.Fetcher != nil {
+		app.HandleFunc("POST /paste", a.handlePaste)
+	}
 	app.HandleFunc("GET /v/{videoId}/status", a.handleStatus)
 	app.HandleFunc("GET /register", a.handleRegisterForm)
 	app.HandleFunc("POST /register", a.handleRegister)
@@ -183,7 +199,7 @@ func (a *App) handleList(w http.ResponseWriter, r *http.Request) {
 	}
 	q := r.URL.Query()
 	f := Filter{
-		Channel:        q.Get("channel"),
+		Channels:       nonEmptyStrings(q["channel"]),
 		From:           q.Get("from"),
 		To:             q.Get("to"),
 		OnlySummarized: q.Get("summarized") == "1",
@@ -198,25 +214,39 @@ func (a *App) handleList(w http.ResponseWriter, r *http.Request) {
 	rows := f.apply(allRows)
 	buckets := bucketRows(rows, a.recencyCutoff())
 
-	// hasConnected drives the empty state: a fresh account with a connection but
+	// Channel options for the multi-select filter (the user's source channels).
-	// no discovery pass yet has zero rows, and we want it to read "connected,
+	channels, err := a.Store.DistinctChannels(r.Context(), userID)
-	// summaries land gradually" rather than "nothing here". Only needed when the
+	if err != nil {
-	// list is empty.
+		a.serverError(w, r, "distinct channels", err)
-	hasConnected := false
+		return
-	if buckets.empty() {
+	}
+
+	// hasConnected drives both the paste box (shown to ANY connected user, #2) and
+	// the empty-state copy (a fresh account with a connection but no discovery pass
+	// yet reads "connected, summaries land gradually" rather than "nothing here").
+	// Computed every render — not only when empty — so a user with videos still
+	// gets the paste box.
 	conns, err := a.Store.ConnectionsForUser(r.Context(), userID)
 	if err != nil {
 		a.serverError(w, r, "connections for user", err)
 		return
 	}
-		hasConnected = len(conns) > 0
+	hasConnected := len(conns) > 0
+
+	// Summarization mode drives the backlog copy: an auto user is told summaries
+	// land gradually; a manual user is told to click Summarize (the first pilot
+	// user sat in manual mode reading "land automatically" and waited forever).
+	autoSummarize, err := a.Store.GetAutoSummarize(r.Context(), userID)
+	if err != nil {
+		a.serverError(w, r, "summarize mode", err)
+		return
 	}
 
 	if isHTMX(r) {
-		a.render(w, r, summaryList(buckets, hasConnected))
+		a.render(w, r, summaryList(buckets, hasConnected, autoSummarize))
 		return
 	}
-	a.render(w, r, ListPage(buckets, f, stats, takeFlash(w, r), hasConnected))
+	a.render(w, r, ListPage(buckets, f, stats, takeFlash(w, r), hasConnected, channels, autoSummarize))
 }
 
 // handleDetail renders one summary in full (highlights, takeaways, action group).
@@ -324,6 +354,75 @@ func (a *App) handleRequestSummarize(w http.ResponseWriter, r *http.Request) {
 	a.render(w, r, VideoCard(*row))
 }
 
+// handlePaste handles "paste a YouTube URL" (Feature 2). It parses the video id,
+// fetches metadata (Data API — ungated), upserts a subscription-less video row
+// scoped to the user (idempotent, so it also dedups), and — if the video isn't
+// already summarized — requests a summary and kicks off immediate processing
+// through the SAME rate gate as the Summarize button. An explicit paste is a
+// manual request, so it summarizes regardless of the recency window. A video that
+// turns out to have no captions resolves to the honest "no transcript" terminal
+// state via the engine (ADR-010), not an error here.
+func (a *App) handlePaste(w http.ResponseWriter, r *http.Request) {
+	userID, ok := a.currentUserID(w, r)
+	if !ok {
+		return
+	}
+	videoID, err := parseYouTubeVideoID(r.FormValue("url"))
+	if err != nil {
+		a.pasteFailure(w, http.StatusBadRequest, "That doesn't look like a YouTube video link.")
+		return
+	}
+
+	v, err := a.Fetcher.FetchVideo(r.Context(), userID, videoID)
+	if errors.Is(err, domain.ErrVideoNotFound) {
+		a.pasteFailure(w, http.StatusNotFound, "That video couldn't be found — it may be private or removed.")
+		return
+	}
+	if err != nil {
+		a.serverError(w, r, "paste fetch", err)
+		return
+	}
+
+	id, err := a.Store.UpsertVideo(r.Context(), v)
+	if err != nil {
+		a.serverError(w, r, "paste upsert", err)
+		return
+	}
+	row, err := a.Store.GetVideoRow(r.Context(), userID, id)
+	if err != nil {
+		a.serverError(w, r, "paste get video", err)
+		return
+	}
+
+	// Dedup: already in the feed with a summary — surface the existing entry,
+	// don't re-summarize.
+	if row.Summarized {
+		a.render(w, r, VideoCard(*row))
+		return
+	}
+
+	// New or unsummarized: queue + (if a Processor is wired) summarize now, through
+	// the shared gate. RequestSummarize makes it durable even if the process dies.
+	if err := a.Store.RequestSummarize(r.Context(), userID, id); err != nil {
+		a.serverError(w, r, "paste request summarize", err)
+		return
+	}
+	if a.Processor != nil {
+		a.startProcessing(userID, id)
+		a.render(w, r, processingCard(*row))
+		return
+	}
+	a.render(w, r, VideoCard(*row))
+}
+
+// pasteFailure renders a minimal inline error fragment for the paste form (HTMX
+// swaps it in). No templ dependency so it renders even on a bad-input fast path.
+func (a *App) pasteFailure(w http.ResponseWriter, status int, msg string) {
+	w.Header().Set("Content-Type", "text/html; charset=utf-8")
+	w.WriteHeader(status)
+	_, _ = io.WriteString(w, `<p class="paste-error" role="alert">`+template.HTMLEscapeString(msg)+`</p>`)
+}
+
 // handleRetryNow handles the "Try now" button on rate-limited video cards. It
 // clears the rate_limited_at backoff so the scheduler won't skip the video, then
 // triggers an immediate ProcessVideo — same background path as handleRequestSummarize.

internal/web/handlers_test.go

@@ -207,13 +207,15 @@ func TestListChannelFilter(t *testing.T) {
 	resetDB(t, p)
 	require.NoError(t, deliver(ctx, app, videoX, "body x"))
 	seedVideo(t, p, videoX, "X Title", "https://x", time.Time{})
+	_, err := p.Exec(ctx, `UPDATE videos SET channel_title = 'Acme Channel' WHERE id = $1`, videoX)
+	require.NoError(t, err)
 
-	// Channel is "youtube" for seeded rows; a non-matching filter hides them.
+	// Selecting a different channel hides the row; selecting its channel shows it.
-	rec := do(t, app, httptest.NewRequest(http.MethodGet, "/?channel=vimeo", nil))
+	rec := do(t, app, httptest.NewRequest(http.MethodGet, "/?channel=Other+Channel", nil))
 	require.Equal(t, http.StatusOK, rec.Code)
 	require.NotContains(t, body(t, rec), "X Title")
 
-	rec = do(t, app, httptest.NewRequest(http.MethodGet, "/?channel=youtube", nil))
+	rec = do(t, app, httptest.NewRequest(http.MethodGet, "/?channel=Acme+Channel", nil))
 	require.Contains(t, body(t, rec), "X Title")
 }
 
@@ -475,3 +477,36 @@ func postAction(t *testing.T, app *web.App, videoID, action string, htmx bool) *
 	}
 	return do(t, app, req)
 }
+
+// TestListManualModeBannerCopy: a manual-mode user with un-summarized videos
+// sees the manual prompt (click Summarize), NOT the "summaries land
+// automatically" copy that misled the first pilot user into waiting forever.
+func TestListManualModeBannerCopy(t *testing.T) {
+	ctx := context.Background()
+	app := newApp(t)
+	p := rawPool(t)
+	resetDB(t, p)
+	seedVideo(t, p, videoX, "X Title", "https://x", time.Time{}) // pending, un-summarized
+	require.NoError(t, app.Store.SetAutoSummarize(ctx, userID, false))
+
+	html := body(t, do(t, app, httptest.NewRequest(http.MethodGet, "/", nil)))
+	require.Contains(t, html, "Manual mode")
+	require.Contains(t, html, "are not summarized automatically")
+	require.NotContains(t, html, "land gradually",
+		"manual-mode user must not be told summaries arrive automatically")
+}
+
+// TestListAutoModeBannerCopy: an auto-mode user with a backlog sees the
+// gradual-delivery copy, not the manual prompt.
+func TestListAutoModeBannerCopy(t *testing.T) {
+	ctx := context.Background()
+	app := newApp(t)
+	p := rawPool(t)
+	resetDB(t, p)
+	seedVideo(t, p, videoX, "X Title", "https://x", time.Time{})
+	require.NoError(t, app.Store.SetAutoSummarize(ctx, userID, true))
+
+	html := body(t, do(t, app, httptest.NewRequest(http.MethodGet, "/", nil)))
+	require.Contains(t, html, "land gradually")
+	require.NotContains(t, html, "are not summarized automatically")
+}

internal/web/paste.go

@@ -0,0 +1,62 @@
+package web
+
+import (
+	"fmt"
+	"net/url"
+	"regexp"
+	"strings"
+)
+
+// youtubeVideoID matches a canonical YouTube video id: exactly 11 URL-safe chars.
+var youtubeVideoID = regexp.MustCompile(`^[A-Za-z0-9_-]{11}$`)
+
+// parseYouTubeVideoID extracts the 11-character video id from a pasted YouTube
+// URL (watch?v=, youtu.be/, shorts/, embed/) or a bare id. It rejects non-YouTube
+// hosts and anything that doesn't yield a valid id, so the paste flow never tries
+// to fetch a video that can't exist (Feature 2).
+func parseYouTubeVideoID(raw string) (string, error) {
+	s := strings.TrimSpace(raw)
+	if s == "" {
+		return "", fmt.Errorf("empty input")
+	}
+
+	// Bare id (no URL) — accept directly.
+	if youtubeVideoID.MatchString(s) {
+		return s, nil
+	}
+
+	// Accept scheme-less URLs (youtube.com/watch?v=...) by giving url.Parse a host.
+	if !strings.Contains(s, "://") {
+		s = "https://" + s
+	}
+	u, err := url.Parse(s)
+	if err != nil {
+		return "", fmt.Errorf("not a URL: %w", err)
+	}
+
+	host := strings.ToLower(u.Hostname())
+	isYouTube := host == "youtu.be" || host == "youtube.com" || strings.HasSuffix(host, ".youtube.com")
+	if !isYouTube {
+		return "", fmt.Errorf("not a YouTube URL: %q", host)
+	}
+
+	var id string
+	switch {
+	case host == "youtu.be":
+		// youtu.be/<id>
+		id = strings.Trim(u.Path, "/")
+	case u.Path == "/watch":
+		id = u.Query().Get("v")
+	default:
+		// /shorts/<id>, /embed/<id>
+		parts := strings.Split(strings.Trim(u.Path, "/"), "/")
+		if len(parts) == 2 && (parts[0] == "shorts" || parts[0] == "embed") {
+			id = parts[1]
+		}
+	}
+
+	if !youtubeVideoID.MatchString(id) {
+		return "", fmt.Errorf("no YouTube video id in %q", raw)
+	}
+	return id, nil
+}

internal/web/paste_handler_test.go

@@ -0,0 +1,134 @@
+package web_test
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+
+	"gitea.d-ma.be/mathias/tapir/internal/domain"
+)
+
+// fakeFetcher is a web.VideoFetcher returning a fixed video (or an error),
+// scoped to whatever (userID, videoID) the handler asks for.
+type fakeFetcher struct {
+	title string
+	err   error
+	calls int
+}
+
+func (f *fakeFetcher) FetchVideo(_ context.Context, userID, videoID string) (domain.Video, error) {
+	f.calls++
+	if f.err != nil {
+		return domain.Video{}, f.err
+	}
+	return domain.Video{
+		UserID:          userID,
+		Provider:        domain.ProviderYouTube,
+		ProviderVideoID: videoID,
+		Title:           f.title,
+		URL:             "https://www.youtube.com/watch?v=" + videoID,
+	}, nil
+}
+
+func pasteReq(rawURL string) *http.Request {
+	req := httptest.NewRequest(http.MethodPost, "/paste",
+		strings.NewReader("url="+url.QueryEscape(rawURL)))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	return req
+}
+
+func TestPasteValidURLAddsAndRequests(t *testing.T) {
+	ctx := context.Background()
+	app := newApp(t)
+	resetDB(t, rawPool(t))
+	app.Fetcher = &fakeFetcher{title: "Pasted Talk"}
+	p := rawPool(t)
+
+	rec := do(t, app, pasteReq("https://youtu.be/dQw4w9WgXcQ"))
+	require.Equal(t, http.StatusOK, rec.Code)
+
+	var (
+		count, requested int
+		title            string
+	)
+	require.NoError(t, p.QueryRow(ctx,
+		`SELECT count(*), coalesce(max(title),'') FROM videos
+		  WHERE user_id=$1 AND provider_video_id='dQw4w9WgXcQ'`, userID).Scan(&count, &title))
+	require.Equal(t, 1, count, "pasted video added once, scoped to the user")
+	require.Equal(t, "Pasted Talk", title)
+
+	require.NoError(t, p.QueryRow(ctx,
+		`SELECT count(*) FROM videos
+		  WHERE user_id=$1 AND provider_video_id='dQw4w9WgXcQ' AND summarize_requested`,
+		userID).Scan(&requested))
+	require.Equal(t, 1, requested, "pasted video is queued for summarization (through the gate)")
+}
+
+func TestPasteInvalidURLRejected(t *testing.T) {
+	app := newApp(t)
+	resetDB(t, rawPool(t))
+	app.Fetcher = &fakeFetcher{title: "x"}
+
+	rec := do(t, app, pasteReq("definitely not a url"))
+	require.Equal(t, http.StatusBadRequest, rec.Code)
+
+	var count int
+	require.NoError(t, rawPool(t).QueryRow(context.Background(),
+		`SELECT count(*) FROM videos WHERE user_id=$1`, userID).Scan(&count))
+	require.Equal(t, 0, count, "invalid input adds nothing")
+}
+
+func TestPasteVideoNotFound(t *testing.T) {
+	app := newApp(t)
+	resetDB(t, rawPool(t))
+	app.Fetcher = &fakeFetcher{err: domain.ErrVideoNotFound}
+
+	rec := do(t, app, pasteReq("https://youtu.be/dQw4w9WgXcQ"))
+	require.Equal(t, http.StatusNotFound, rec.Code)
+
+	var count int
+	require.NoError(t, rawPool(t).QueryRow(context.Background(),
+		`SELECT count(*) FROM videos WHERE user_id=$1`, userID).Scan(&count))
+	require.Equal(t, 0, count, "a not-found video adds nothing")
+}
+
+func TestPasteDedupNoDuplicate(t *testing.T) {
+	app := newApp(t)
+	resetDB(t, rawPool(t))
+	app.Fetcher = &fakeFetcher{title: "Pasted Talk"}
+
+	require.Equal(t, http.StatusOK, do(t, app, pasteReq("https://youtu.be/dQw4w9WgXcQ")).Code)
+	require.Equal(t, http.StatusOK, do(t, app, pasteReq("https://www.youtube.com/watch?v=dQw4w9WgXcQ")).Code)
+
+	var count int
+	require.NoError(t, rawPool(t).QueryRow(context.Background(),
+		`SELECT count(*) FROM videos WHERE user_id=$1 AND provider_video_id='dQw4w9WgXcQ'`,
+		userID).Scan(&count))
+	require.Equal(t, 1, count, "pasting the same video twice must not duplicate the row")
+}
+
+func TestListShowsPasteFormForConnectedUserWithVideos(t *testing.T) {
+	app := newApp(t)
+	resetDB(t, rawPool(t))
+	app.Fetcher = &fakeFetcher{title: "x"}
+	p := rawPool(t)
+
+	// Connected user with a non-empty feed (the case the bug missed: hasConnected
+	// was only computed for an empty feed).
+	_, err := p.Exec(context.Background(),
+		`INSERT INTO video_connections (user_id, provider, token_ref, status)
+		 VALUES ($1, 'youtube', 'youtube/x/refresh_token', 'active')`, userID)
+	require.NoError(t, err)
+	seedVideo(t, p, "11111111-1111-1111-1111-111111111111", "A talk", "https://youtu.be/aaaaaaaaaaa", time.Now())
+
+	rec := do(t, app, httptest.NewRequest(http.MethodGet, "/", nil))
+	require.Equal(t, http.StatusOK, rec.Code)
+	require.Contains(t, body(t, rec), `action="/paste"`,
+		"a connected user must see the paste box even when the feed has videos")
+}

internal/web/paste_test.go

@@ -0,0 +1,97 @@
+package web
+
+import (
+	"bytes"
+	"context"
+	"gitea.d-ma.be/mathias/tapir/internal/adapters/store"
+	"strings"
+	"testing"
+)
+
+func TestParseYouTubeVideoID(t *testing.T) {
+	const id = "dQw4w9WgXcQ"
+	ok := []struct {
+		name, in string
+	}{
+		{"watch", "https://www.youtube.com/watch?v=" + id},
+		{"watch no www", "https://youtube.com/watch?v=" + id},
+		{"watch m", "https://m.youtube.com/watch?v=" + id},
+		{"watch extra params", "https://www.youtube.com/watch?v=" + id + "&t=42s&list=PLxyz"},
+		{"watch param after", "https://www.youtube.com/watch?list=PLxyz&v=" + id},
+		{"short link", "https://youtu.be/" + id},
+		{"short link param", "https://youtu.be/" + id + "?si=abcd&t=1"},
+		{"shorts", "https://www.youtube.com/shorts/" + id},
+		{"embed", "https://www.youtube.com/embed/" + id},
+		{"bare id", id},
+		{"http scheme", "http://youtube.com/watch?v=" + id},
+		{"no scheme", "youtube.com/watch?v=" + id},
+		{"trailing space", "  https://youtu.be/" + id + "  "},
+	}
+	for _, c := range ok {
+		t.Run(c.name, func(t *testing.T) {
+			got, err := parseYouTubeVideoID(c.in)
+			if err != nil {
+				t.Fatalf("parseYouTubeVideoID(%q) error: %v", c.in, err)
+			}
+			if got != id {
+				t.Fatalf("parseYouTubeVideoID(%q) = %q, want %q", c.in, got, id)
+			}
+		})
+	}
+
+	bad := []struct {
+		name, in string
+	}{
+		{"empty", ""},
+		{"blank", "   "},
+		{"vimeo", "https://vimeo.com/123456789"},
+		{"other host", "https://example.com/watch?v=" + id},
+		{"watch no id", "https://www.youtube.com/watch?v="},
+		{"short id", "https://youtu.be/abc"},
+		{"long id", "https://youtu.be/" + id + "extra"},
+		{"bad chars", "https://youtu.be/dQw4w9Wg!cQ"},
+		{"not a url", "just some text"},
+		{"channel url", "https://www.youtube.com/@somechannel"},
+	}
+	for _, c := range bad {
+		t.Run("reject "+c.name, func(t *testing.T) {
+			if got, err := parseYouTubeVideoID(c.in); err == nil {
+				t.Fatalf("parseYouTubeVideoID(%q) = %q, want error", c.in, got)
+			}
+		})
+	}
+}
+
+func TestListPageShowsPasteFormOnlyWhenConnected(t *testing.T) {
+	render := func(connected bool) string {
+		var buf bytes.Buffer
+		if err := ListPage(listBuckets{}, Filter{}, PipelineStats{}, "", connected, nil, true).Render(context.Background(), &buf); err != nil {
+			t.Fatalf("render: %v", err)
+		}
+		return buf.String()
+	}
+
+	html := render(true)
+	if !strings.Contains(html, `name="url"`) || !strings.Contains(html, `action="/paste"`) {
+		t.Errorf("connected feed must show the paste form")
+	}
+	if strings.Contains(render(false), `name="url"`) {
+		t.Errorf("disconnected feed must not show the paste form")
+	}
+}
+
+func TestFilterMatchesMultipleChannels(t *testing.T) {
+	f := Filter{Channels: []string{"Acme", "Zeta"}}
+	row := func(ch string) store.SummaryRow { return store.SummaryRow{ChannelTitle: ch, Summarized: true} }
+
+	rows := []store.SummaryRow{row("Acme"), row("Beta"), row("Zeta")}
+	got := f.apply(rows)
+	if len(got) != 2 || got[0].ChannelTitle != "Acme" || got[1].ChannelTitle != "Zeta" {
+		t.Fatalf("multi-channel filter = %+v, want Acme+Zeta only", got)
+	}
+
+	// Empty selection = no channel constraint (all pass).
+	if n := len(Filter{}.apply(rows)); n != 3 {
+		t.Fatalf("no channel filter should pass all rows, got %d", n)
+	}
+}

internal/web/processing.go

@@ -3,6 +3,8 @@ package web
 import (
 	"context"
 	"sync"
+
+	"gitea.d-ma.be/mathias/tapir/internal/domain"
 )
 
 // Processor runs the core summarization use case for a single already-discovered
@@ -14,6 +16,14 @@ type Processor interface {
 	ProcessVideo(ctx context.Context, userID, videoID string) error
 }
 
+// VideoFetcher resolves an arbitrary YouTube video id to its metadata for the
+// paste-a-URL flow (Feature 2). It is a Data API call, NOT the rate-limited
+// caption path. Returns domain.ErrVideoNotFound for a deleted/private/typo'd id.
+// cmd/tapir wires a per-user YouTube adapter; nil disables the paste route.
+type VideoFetcher interface {
+	FetchVideo(ctx context.Context, userID, videoID string) (domain.Video, error)
+}
+
 // ProcessingSet tracks the (user, video) ids currently being summarized in-process
 // so the status endpoint can show the animation until the summary lands. It is
 // ephemeral (single-instance Stage-1): a restart drops it, and the DB holds the

internal/web/view.go

@@ -2,6 +2,7 @@ package web
 
 import (
 	"regexp"
+	"slices"
 	"strings"
 	"time"
 	"unicode/utf8"
@@ -333,7 +334,7 @@ type flashView struct {
 // flashMessages maps each flash code to its banner. An unknown code renders no
 // banner (flashFor returns ok=false), so a forged cookie value is inert.
 var flashMessages = map[string]flashView{
-	flashConnected:     {"success", "YouTube account connected."},
+	flashConnected:     {"success", "YouTube account connected — finding your subscriptions. Your newest videos will appear below as they're summarized."},
 	flashConnectFailed: {"error", "Could not connect your YouTube account. Please try again."},
 	flashDisconnected:  {"success", "Account disconnected."},
 	flashDeleted:       {"success", "Your account and all its data were deleted."},
@@ -469,7 +470,7 @@ func (b listBuckets) empty() bool {
 // Dates are kept as the raw YYYY-MM-DD strings so the form re-renders the user's
 // input verbatim; parsing happens in matchFilter.
 type Filter struct {
-	Channel        string
+	Channels       []string // selected channel titles; empty = all channels
 	From           string
 	To             string
 	OnlySummarized bool // show only videos that have a summary
@@ -480,7 +481,28 @@ type Filter struct {
 // filter) the bar is hidden so the connect CTA stands alone (UX review C1); a
 // filter that happens to match nothing still shows the bar so it can be cleared.
 func (f Filter) active() bool {
-	return f.Channel != "" || f.From != "" || f.To != "" || f.OnlySummarized
+	return len(f.Channels) > 0 || f.From != "" || f.To != "" || f.OnlySummarized
+}
+
+// HasChannel reports whether a channel is currently selected (drives the
+// multi-select's selected state in the view).
+func (f Filter) HasChannel(c string) bool {
+	return slices.Contains(f.Channels, c)
+}
+
+// nonEmptyStrings drops blank entries. A channel multi-select submits real
+// channel titles; this guards against a stray empty value reaching the filter.
+func nonEmptyStrings(ss []string) []string {
+	out := ss[:0:0]
+	for _, s := range ss {
+		if strings.TrimSpace(s) != "" {
+			out = append(out, s)
+		}
+	}
+	if len(out) == 0 {
+		return nil
+	}
+	return out
 }
 
 // matches reports whether a row satisfies the filter. Channel is an exact match;
@@ -491,7 +513,7 @@ func (f Filter) matches(r store.SummaryRow) bool {
 	if f.OnlySummarized && !r.Summarized {
 		return false
 	}
-	if f.Channel != "" && r.Channel != f.Channel {
+	if len(f.Channels) > 0 && !slices.Contains(f.Channels, r.ChannelTitle) {
 		return false
 	}
 	if from, ok := parseDate(f.From); ok {
@@ -521,7 +543,7 @@ func parseDate(s string) (time.Time, bool) {
 
 // apply returns the subset of rows matching the filter, preserving order.
 func (f Filter) apply(rows []store.SummaryRow) []store.SummaryRow {
-	if f.Channel == "" && f.From == "" && f.To == "" && !f.OnlySummarized {
+	if len(f.Channels) == 0 && f.From == "" && f.To == "" && !f.OnlySummarized {
 		return rows
 	}
 	out := rows[:0:0]

internal/web/views.templ

@@ -104,23 +104,35 @@ templ flashBanner(code string) {
 // #summary-list region; a non-HTMX request renders the whole page. flash carries
 // a one-shot notification (e.g. "connected", "registered") surfaced on arrival
 // after a POST→redirect.
-templ ListPage(b listBuckets, f Filter, stats PipelineStats, flash string, hasConnected bool) {
+templ ListPage(b listBuckets, f Filter, stats PipelineStats, flash string, hasConnected bool, channels []string, autoSummarize bool) {
 	@Layout("Tapir — Summaries") {
 		@flashBanner(flash)
+		if hasConnected {
+			@pasteForm()
+		}
 		if !b.empty() || f.active() {
-			@filterForm(f)
+			@filterForm(f, channels)
 		}
 		if stats.RateLimited > 0 || stats.Pending > 0 || stats.NoText > 0 {
 			@pipelineBar(stats)
 		}
-		if stats.RateLimited+stats.Pending > 0 {
+		if (stats.RateLimited+stats.Pending) > 0 && autoSummarize {
 			<p class="pipeline-note muted">
 				Tapir fetches captions slowly on purpose, to respect YouTube's limits —
 				new summaries land gradually. Check back tomorrow.
 			</p>
 		}
+		if (stats.RateLimited+stats.Pending) > 0 && !autoSummarize {
+			<p class="pipeline-note muted">
+				You are in Manual mode: new videos appear here but are not summarized
+				automatically. Use the Summarize button on the ones you want.
+			</p>
+			<p class="pipeline-note muted">
+				<a href="/account">Switch to Automatic</a> to have new videos summarized for you.
+			</p>
+		}
 		<div id="summary-list">
-			@summaryList(b, hasConnected)
+			@summaryList(b, hasConnected, autoSummarize)
 		</div>
 	}
 }
@@ -143,7 +155,29 @@ templ pipelineBar(s PipelineStats) {
 	</div>
 }
 
-templ filterForm(f Filter) {
+// pasteForm lets a connected user summarize any YouTube video by pasting its URL
+// (Feature 2). The result (a video card, or an inline error) swaps into
+// #paste-result; the next list refresh shows it inline. Summarization runs
+// through the shared caption rate gate like every other fetch.
+templ pasteForm() {
+	<form
+		class="paste"
+		method="post"
+		action="/paste"
+		hx-post="/paste"
+		hx-target="#paste-result"
+		hx-swap="innerHTML"
+	>
+		<label>
+			Summarize any video
+			<input type="url" name="url" placeholder="Paste a YouTube link…" required/>
+		</label>
+		<button type="submit">Add</button>
+	</form>
+	<div id="paste-result"></div>
+}
+
+templ filterForm(f Filter, channels []string) {
 	<form
 		class="filters"
 		method="get"
@@ -153,7 +187,16 @@ templ filterForm(f Filter) {
 		hx-swap="innerHTML"
 		hx-indicator="#filter-indicator"
 	>
-		<label>Channel <input type="text" name="channel" value={ f.Channel } placeholder="any"/></label>
+		if len(channels) > 0 {
+			<label>
+				Channels
+				<select name="channel" multiple size="4">
+					for _, c := range channels {
+						<option value={ c } selected?={ f.HasChannel(c) }>{ c }</option>
+					}
+				</select>
+			</label>
+		}
 		<label class="filter-check">
 			<input type="checkbox" name="summarized" value="1" if f.OnlySummarized { checked }/>
 			Summarized only
@@ -169,12 +212,16 @@ templ filterForm(f Filter) {
 // and a single disclosure holding the older un-summarized back-catalogue. Cards
 // reflow to a single column on mobile; an empty list shows a friendly first-run
 // state instead of a blank table.
-templ summaryList(b listBuckets, hasConnected bool) {
+templ summaryList(b listBuckets, hasConnected bool, autoSummarize bool) {
 	if b.empty() {
 		if hasConnected {
 			<div class="empty empty-connected">
 				<strong>Your account is connected</strong>
+				if autoSummarize {
 					<span>Tapir is finding your subscriptions and fetching captions — summaries appear here gradually. Check back later.</span>
+				} else {
+					<span>Tapir is finding your subscriptions. You are in Manual mode, so videos appear here with a Summarize button — pick the ones you want, or switch to Automatic in your account.</span>
+				}
 			</div>
 		} else {
 			<div class="empty">

test/acceptance/scenario_coverage_test.go

@@ -0,0 +1,221 @@
+package acceptance
+
+// This is the name-coverage gate for the BDD spec (see docs/use-cases/*.feature).
+// There is no godog runner — the .feature files are design records, and the real
+// behaviour is covered by the hand-written Go tests across the module. This test
+// keeps the two from drifting in the cheapest honest way: every non-@pending
+// Scenario must have an entry in scenarioCoverage pointing at a Go test that
+// actually exists. It does NOT prove the test exercises the scenario (only godog
+// could); it catches the common drift — "added a scenario, forgot the test", a
+// renamed/deleted covering test, or a scenario removed without cleaning the map.
+//
+// When you add a Scenario: either map it here to its covering test, or tag it
+// @pending in the .feature with a one-line reason for why it has no test yet.
+
+import (
+	"os"
+	"path/filepath"
+	"regexp"
+	"strings"
+	"testing"
+)
+
+// scenarioCoverage maps each non-@pending Scenario name to the Go test that
+// covers it. Keep it in sync with docs/use-cases/*.feature — the test below
+// fails if a scenario is unmapped, a mapped test is missing, or an entry no
+// longer matches a real non-pending scenario.
+var scenarioCoverage = map[string]string{
+	// ai_routing.feature
+	"Local AI produces the summary":                                     "TestSummarize_LocalSucceeds",
+	"Local AI fails and the user has a BYO provider configured":         "TestSummarize_FallsBackToBYO",
+	"Local AI fails and the user has no BYO provider":                   "TestSummarize_LocalFailsNoBYO_NoExternalSend",
+	"A user without BYO never has content sent externally":              "TestSummarize_NoBYO_ContentOnlyLocal",
+	"A model returns unparseable output and the next endpoint succeeds": "TestSummarize_FallsBackOnMalformedOutput",
+
+	// landing_page.feature
+	"An unauthenticated visit to the root is sent to the welcome page":    "TestUnauthenticatedRootRedirectsToWelcome",
+	"The welcome page invites an unauthenticated visitor to start":        "TestWelcomeLoggedOut",
+	"An authenticated user on the welcome page sees their way in and out": "TestWelcomeLoggedIn",
+
+	// connect_account.feature
+	"Connect a YouTube account":                          "TestCallbackExchangesAndRecordsConnection",
+	"Connecting an account discovers videos immediately": "TestCallbackTriggersDiscovery",
+	"Connecting summarizes my newest videos right away":  "TestNewestUnsummarizedVideoIDs",
+	"Tokens are never stored in the clear":               "TestCallbackExchangesAndRecordsConnection",
+
+	// paste_url.feature
+	"Paste a valid YouTube URL":                              "TestPasteValidURLAddsAndRequests",
+	"Pasting an invalid link is rejected":                    "TestPasteInvalidURLRejected",
+	"Pasting a video that cannot be found is honest":         "TestPasteVideoNotFound",
+	"Pasting the same video twice does not duplicate it":     "TestPasteDedupNoDuplicate",
+	"Revoking a connection stops watching but keeps history": "TestDisconnectRemovesTokenAndConnectionKeepsAccount",
+
+	// summarize_mode.feature
+	"Auto mode summarizes recent new videos automatically":           "TestRunOnce_AutoMode_SkipsOldVideos",
+	"Auto mode lists older videos without summarizing them":          "TestRunOnce_AutoMode_SkipsOldVideos",
+	"Automatic is the default for a new user":                        "TestRegisteredUserDefaultsAutoSummarizeOn",
+	"Manual mode leaves new videos unsummarized":                     "TestRunOnce_ManualMode_SkipsUnrequested",
+	"Requesting a summary in manual mode queues it for the next run": "TestRunOnce_ManualMode_ProcessesRequested",
+
+	// registration.feature
+	"A new Dex subject is routed to registration":                               "TestUnregisteredSubjectRedirectedToRegister",
+	"Registering creates the account and its identity mapping":                  "TestRegisterCreatesExactlyOneUserAndIdentity",
+	"A returning subject passes straight through":                               "TestRegisteredSubjectPassesThrough",
+	"Deleting an account removes only my data and leaves other users untouched": "TestDeleteAccountWipesDataAndSecretsAndLogsOut",
+
+	// summarize_new_video.feature
+	"A subscribed channel posts a video that has captions":         "TestSubscribedVideoWithCaptionsIsSummarizedAndDelivered",
+	"A subscribed channel posts a video with no usable transcript": "TestVideoWithNoTranscriptIsSkipped",
+	"A channel I am not subscribed to posts a video":               "TestUnsubscribedChannelVideoIsNotProcessed",
+	"The same video is not summarized twice":                       "TestAlreadySummarizedVideoIsNotReprocessed",
+	"Re-analyzing a stored video does not re-fetch its transcript": "TestProcessNewVideo_SecondSummarizeDoesNotRefetch",
+}
+
+var (
+	scenarioRe = regexp.MustCompile(`^\s*Scenario(?: Outline)?:\s*(.+?)\s*$`)
+	testFuncRe = regexp.MustCompile(`^func (Test\w+)\(`)
+)
+
+// scenario is one parsed Gherkin scenario and whether it is @pending.
+type scenario struct {
+	name    string
+	pending bool
+}
+
+func TestScenarioCoverage(t *testing.T) {
+	root := moduleRoot(t)
+
+	scenarios := parseScenarios(t, filepath.Join(root, "docs", "use-cases"))
+	if len(scenarios) == 0 {
+		t.Fatal("no scenarios parsed from docs/use-cases — wrong path?")
+	}
+	tests := allTestFuncNames(t, root)
+
+	// Index scenario names for the reverse (stale-entry) check.
+	active := map[string]bool{} // non-pending scenario names
+	var pending []string
+	for _, s := range scenarios {
+		if s.pending {
+			pending = append(pending, s.name)
+			continue
+		}
+		active[s.name] = true
+
+		// 1. Every non-pending scenario must be mapped.
+		fn, ok := scenarioCoverage[s.name]
+		if !ok {
+			t.Errorf("scenario %q has no coverage entry — map it in scenarioCoverage to a covering test, or tag it @pending in the .feature", s.name)
+			continue
+		}
+		// 2. The mapped test must actually exist.
+		if !tests[fn] {
+			t.Errorf("scenario %q maps to %q, which is not a Test function anywhere in the module", s.name, fn)
+		}
+	}
+
+	// 3. No stale entries: every map key must be a real, non-pending scenario.
+	for name := range scenarioCoverage {
+		if !active[name] {
+			t.Errorf("scenarioCoverage has entry %q, which is not a current non-pending scenario (renamed, removed, or now @pending?)", name)
+		}
+	}
+
+	if len(pending) > 0 {
+		t.Logf("%d @pending scenario(s) without a test (tracked, not required): %s",
+			len(pending), strings.Join(pending, "; "))
+	}
+}
+
+// parseScenarios reads every *.feature under dir and returns its scenarios with
+// their @pending status. A scenario is @pending when a `@pending` tag line
+// precedes it (tags survive intervening comment lines, the layout these files
+// use); the flag is consumed at the Scenario line and reset afterward.
+func parseScenarios(t *testing.T, dir string) []scenario {
+	t.Helper()
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		t.Fatalf("read use-cases dir: %v", err)
+	}
+	var out []scenario
+	for _, e := range entries {
+		if e.IsDir() || !strings.HasSuffix(e.Name(), ".feature") {
+			continue
+		}
+		b, err := os.ReadFile(filepath.Join(dir, e.Name()))
+		if err != nil {
+			t.Fatalf("read %s: %v", e.Name(), err)
+		}
+		pending := false
+		for _, line := range strings.Split(string(b), "\n") {
+			trimmed := strings.TrimSpace(line)
+			if strings.HasPrefix(trimmed, "@") {
+				if strings.Contains(trimmed, "@pending") {
+					pending = true
+				}
+				continue
+			}
+			if m := scenarioRe.FindStringSubmatch(line); m != nil {
+				out = append(out, scenario{name: m[1], pending: pending})
+				pending = false
+			}
+			// comment (#) and step lines leave a set @pending intact until the
+			// scenario consumes it; a blank line between scenarios is harmless.
+		}
+	}
+	return out
+}
+
+// allTestFuncNames walks the module for `func TestXxx(` declarations, excluding
+// this file (whose regex literal would otherwise look like a definition).
+func allTestFuncNames(t *testing.T, root string) map[string]bool {
+	t.Helper()
+	self := "scenario_coverage_test.go"
+	names := map[string]bool{}
+	err := filepath.WalkDir(root, func(path string, d os.DirEntry, err error) error {
+		if err != nil {
+			return err
+		}
+		if d.IsDir() {
+			if d.Name() == ".git" {
+				return filepath.SkipDir
+			}
+			return nil
+		}
+		if !strings.HasSuffix(path, "_test.go") || filepath.Base(path) == self {
+			return nil
+		}
+		b, err := os.ReadFile(path)
+		if err != nil {
+			return err
+		}
+		for _, line := range strings.Split(string(b), "\n") {
+			if m := testFuncRe.FindStringSubmatch(line); m != nil {
+				names[m[1]] = true
+			}
+		}
+		return nil
+	})
+	if err != nil {
+		t.Fatalf("walk module: %v", err)
+	}
+	return names
+}
+
+// moduleRoot walks up from the working directory to the dir containing go.mod.
+func moduleRoot(t *testing.T) string {
+	t.Helper()
+	dir, err := os.Getwd()
+	if err != nil {
+		t.Fatalf("getwd: %v", err)
+	}
+	for {
+		if _, err := os.Stat(filepath.Join(dir, "go.mod")); err == nil {
+			return dir
+		}
+		parent := filepath.Dir(dir)
+		if parent == dir {
+			t.Fatal("go.mod not found walking up from cwd")
+		}
+		dir = parent
+	}
+}