# tapir

Watches a user's YouTube/Vimeo subscriptions and, when a subscribed channel posts a new
video, summarizes it into highlights and takeaways using a **local-first** AI stack with an
optional, per-user **BYO-AI** fallback. **Standalone-first**; feeding a personal knowledge
base ("brain") is one optional sink, not the reason Tapir exists. Written in Go.

## Status

**Stage-0 demo slice built.** The engine, ports, domain, and adapters (`llm`, `summarizer`,
captions-first `youtube`, Postgres `store`) are implemented and green, with the `tapir`
CLI (`auth` / `run` / `list` / `show`) wiring the end-to-end loop for a single user. The
remaining work to *open* the Stage-0 clock is the live run on real subscriptions — see
"Running the Stage-0 demo" below. The guardrail docs (vision, decisions, architecture, data
model, behavior specs) remain the source of intent.

## Read these first (the guardrails)

| Doc | What it is |
|-----|-----------|
| [`VISION.md`](VISION.md) | Product vision, principles, and the **staged Definition of Success**. Stage 0 ("useful to me") is the gate before any multi-user work. |
| [`DECISIONS.md`](DECISIONS.md) | Architecture Decision Records (append-only). Why Go, why no Supabase, standalone-first, captions-first, etc. |
| [`docs/architecture/architecture.md`](docs/architecture/architecture.md) | C4 context + container diagrams, key sequence diagrams, and the Clean Architecture layering (Mermaid). |
| [`docs/data-model.md`](docs/data-model.md) | Entities and the per-user isolation model (Stage 0 / Stage 1 scope). |
| [`docs/use-cases/`](docs/use-cases/) | **Gherkin `.feature` files** — the BDD behavior spec that seeds the test suite. |

## Approach

- **Clean Architecture / ports & adapters.** A provider- and sink-agnostic engine depends only
  on interfaces (`VideoSource`, `Summarizer`, `Sink`, `SecretStore`). YouTube, Vimeo, the AI
  router, the user store, and the brain sink are adapters. "Standalone vs homelab" is a wiring
  choice, not two codebases.
- **TDD/BDD.** The `.feature` files are the living behavior spec; the use-case core is tested
  through fake adapters. Behavior is specified as executable scenarios, not prose that drifts.
- **Trunk-Based Development.** Commit directly to `main`, one logical change per commit, every
  commit deployable (see ADR-009). CI is the quality gate.

## Running the Stage-0 demo

Tapir runs on your YouTube account(s): authorize once, then run the
watch→summarize→deliver loop. All configuration is via `TAPIR_*` environment
variables — copy [`.env.example`](.env.example) to `.env` and fill it in (no
secrets are committed; at demo time source them from op, e.g. `op run -- ...`).

```sh
# 1. configure (UUID user id, gateway URL+key, Postgres DSN, YouTube OAuth app,
#    summarizer model). See .env.example for every variable.
cp .env.example .env && $EDITOR .env
set -a && . ./.env && set +a          # export them into the shell

go build -o bin/tapir ./cmd/tapir

# 2. one-time: authorize YouTube. Opens a consent URL, captures the redirect on
#    TAPIR_OAUTH_REDIRECT_ADDR, and stores the refresh token via the SecretStore
#    (a 0600 file at Stage 0). The token is never logged.
./bin/tapir auth

# 3. run: detect new videos across your subscriptions, summarize, deliver to the
#    store. Single pass; set TAPIR_DISCOVERY_INTERVAL (e.g. 2h) for the serve loop.
./bin/tapir run
```

Live prerequisites at demo time: the LiteLLM gateway reachable
(`TAPIR_GATEWAY_URL` + a valid key — resolve from op, the documented
`sk-local-123` is stale), a Postgres DSN (`TAPIR_DB_DSN`, migrations apply on
first connect), and a registered YouTube OAuth client whose authorized redirect
URI matches `TAPIR_OAUTH_REDIRECT_ADDR`. The summarizer model
(`TAPIR_SUMMARIZER_MODEL`, default `koala/phi4-mini`) is overridable; pick the
final alias when the gateway is reachable (see `docs/homelab-integration.md`).

### Web surface (`tapir serve`)

`tapir serve` starts the HTMX+Templ web UI on `:8080`. Users log in via Dex OIDC (local
password or Google); a new Dex subject is routed to `/register` to create a Tapir account.
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. In auto mode only videos published within
`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

koala has no browser and no interactive `op` session, so the two interactive
edges are handled without changing any code:

**Secrets via an `op` service account** (no `op signin`). Create a 1Password
service account with read on the `HomeLab` vault, export its token, and keep one
op-style env file `tapir.env` — secret values as `op://` refs, the rest as
literals — then prefix every command with `op run`:

```sh
export OP_SERVICE_ACCOUNT_TOKEN=ops_...        # the service-account token

cat > tapir.env <<'EOF'
TAPIR_USER_ID=<your-uuid>
TAPIR_GATEWAY_URL=http://koala:30401/v1
TAPIR_GATEWAY_KEY=op://HomeLab/LITELLM_MASTER_KEY/password
TAPIR_SUMMARIZER_MODEL=koala/phi4-mini
TAPIR_DB_DSN=op://HomeLab/TAPIR_DB_DSN/password
TAPIR_YT_CLIENT_ID=op://HomeLab/TAPIR_YT_OAUTH/client_id
TAPIR_YT_CLIENT_SECRET=op://HomeLab/TAPIR_YT_OAUTH/client_secret
TAPIR_YT_TOKEN_REF=youtube/refresh_token
TAPIR_SECRETS_FILE=/home/mathias/.config/tapir/secrets.json
TAPIR_OAUTH_REDIRECT_ADDR=localhost:8080
EOF
chmod 600 tapir.env
```

(Create the `TAPIR_DB_DSN` and `TAPIR_YT_OAUTH` items in the `HomeLab` vault
first; `LITELLM_MASTER_KEY` already exists. `sk-local-123` is stale.)

**`tapir auth` over an SSH tunnel.** The auth command binds a listener on
`localhost:8080` *on koala* and prints the consent URL to stdout — it never
opens a browser. Forward that port to your laptop, run auth, and approve in your
laptop's browser; the redirect returns through the tunnel:

```sh
# from your laptop:
ssh -L 8080:localhost:8080 koala
# now on koala (one-time):
op run --env-file tapir.env -- ./bin/tapir auth
#   → copy the printed URL into your laptop browser, approve; token is stored.
```

The Google OAuth client must be a **Desktop/Web** type with authorized redirect
`http://localhost:8080/callback` (matching `TAPIR_OAUTH_REDIRECT_ADDR`), the
YouTube Data API enabled, and your account added as a test user.

**Everything else is already non-interactive** — once the token is stored, run
headless:

```sh
op run --env-file tapir.env -- ./bin/tapir run     # detect → summarize → deliver
op run --env-file tapir.env -- ./bin/tapir list     # read summaries back
op run --env-file tapir.env -- ./bin/tapir show <video-id>
```

Postgres (`postgres18`) is ClusterIP-only; from the koala host reach it with
`kubectl port-forward -n databases svc/postgres18 5432:5432` and point
`TAPIR_DB_DSN` at `localhost:5432` (migrations apply on first connect).

## Conventions

Reuses homelab conventions: Go, Dex for identity, ESO + 1Password for secrets, Postgres for
persistence. No new auth or secrets system (ADR-002).

## Next

First implementation step: scaffold the Go service (engine + interfaces + the copied `llm`
package), captions-first, with the store and brain sink adapters — decomposable into
independent units suitable for a Claude Code swarm. Tracked as the first build issue.