-
feat: per-video chat over the stored transcript (ADR-027)
released this
2026-06-11 07:57:24 +00:00 | 34 commits to main since this releaseDeeper-dive chat entered from the summary view, answering questions against a
video's already-stored transcript (ADR-021). Safe by construction: no caption
fetch, no rate gate, no YouTube. Model defaults to the summary's model and is
switchable among the ADR-022 chain; cloud model absent when disabled. Ephemeral
v1 — no persisted history, no migration.Downloads