summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorbot <bot@x>2026-11-13 11:39:28 +0000
committerbot <bot@x>2026-11-13 11:39:28 +0000
commit99a0a5e3278f1d43b7ddfc99dd8e98ceaf8067ad (patch)
tree751eedaebfcd85a4c9f5fb64d3d3954e7a3ae12c
parent5863806dd4e15c63174bee4e7515a4256792ecfa (diff)
docs(agents): animatic toolchain config, and every PR carries an animaticanimatic-config
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
-rw-r--r--CLAUDE.md6
-rw-r--r--docs/agents/animatic.md26
2 files changed, 32 insertions, 0 deletions
diff --git a/CLAUDE.md b/CLAUDE.md
index cc3dbb3..7c571b6 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -31,3 +31,9 @@ Every skill that dispatches a subagent says to start it with a fresh context win
### Triage labels
Canonical names, unchanged. See `docs/agents/triage-labels.md`. Issues are judged against [`SCOPE.md`](./SCOPE.md).
+
+### Animatic
+
+The animatic skill's voice, renderer, style and stitcher. See `docs/agents/animatic.md`.
+
+Every PR carries an animatic. Once the PR is open, run the `animatic` skill on it and post the video as a PR comment with `gh pr comment <number> --attach <file>.mp4`.
diff --git a/docs/agents/animatic.md b/docs/agents/animatic.md
new file mode 100644
index 0000000..60e1203
--- /dev/null
+++ b/docs/agents/animatic.md
@@ -0,0 +1,26 @@
+# Animatic
+
+The `animatic` skill's toolchain for this repo, chosen and tested with `setup-animatic`. Edit it by hand, or run `setup-animatic` again to change a tool.
+
+## Flags
+
+Pass these, unchanged, to every `animatic.mjs` command (`build`, `say`, `setup`). Speech is cached per voice flag, so a run with different flags re-voices every line.
+
+```
+--tts-command 'edge-tts --voice en-US-AndrewNeural --text {text} --write-media {out}.mp3'
+```
+
+## Stages
+
+- **Voice**: edge-tts, `en-US-AndrewNeural` (Microsoft's neural voices, online, no key). Needs `edge-tts` on `PATH` (`pipx install edge-tts`).
+- **Renderer**: generated frames on the kit (Playwright).
+- **Style**: the plain kit.
+- **Stitcher**: ffmpeg, through `build`.
+
+## Secrets
+
+None.
+
+## Tested
+
+2026-11-13: the voice at 2.89 words/s (from `say`); a one-moment build verified with no warnings.