mirror of
https://github.com/digitalsamba/claude-code-video-toolkit.git
synced 2026-09-18 19:41:13 +08:00
Conal Mullan
290d7d345b
* Adopt uv for Python dependency management Replace manual pip/venv setup with `uv sync` (pyproject.toml + uv.lock), and switch all tool invocations across docs, skills, commands, and examples to `uv run`. Optional extras: --extra whisper, --extra modal. tools/requirements.txt is kept as a pip fallback. * FIX: bring uv migration current with main — post-June deps and docs Adds websocket-client (60db websocket transport) to core deps and a 'youtube' extra for the google-* packages (tools/youtube_upload.py), regenerates uv.lock, bumps pyproject version to 0.18.0, and converts the 51 'python3 tools/…' invocations added since the PR was opened (60db, YouTube, Kiro docs, publish command) to 'uv run'. Verified with uv sync --frozen --extra youtube; voiceover, sixtydb_tts, youtube_upload and migrate_to_kiro all run through the project environment. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Asharib Ali <asharibarain4@gmail.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
sky-blue-short — concept-explainer-short showcase
A complete worked example of the concept-explainer-short template: a 52-second 9:16 vertical short answering "Why is the sky blue?" — every asset AI-generated, total cloud cost ≈ $0.80.
All source assets are committed, so you can study or rebuild it immediately:
cd examples/sky-blue-short
uv run build.py # re-render from committed assets → out/short.mp4
How it was made
| Piece | Tool | Notes |
|---|---|---|
| 4-scene plan + narration | scenes.json |
hook → concept card → payoff → CTA |
| Voiceover | gen_vo.py → Qwen3-TTS (Modal) |
built-in speaker Ryan, tone: warm, one batch call; pacing QC reported 125–152 wpm, no clamping needed |
| Sky + sunset b-roll | tools/ltx2.py |
576×1024, 161 frames, seed 11 — prompts in the table below |
| Scattering diagram + CTA cards | tools/ideogram4.py |
JSON captions in captions/02_scattering.json / 04_cta.json, --resolution 1440x2560 |
| Music bed | tools/music_gen.py |
60s ambient, looped + ducked by build.py |
| Caption timing | gen_captions.py |
whisper word timestamps force-aligned to the script |
LTX prompts used:
01_hook— "Low angle looking straight up at a brilliant deep blue sky, wispy white clouds drifting slowly overhead, sun flaring at the edge of the frame with soft sunbeams, peaceful atmosphere, cinematic, vertical composition"03_sunset— "A glowing orange and red sunset over the ocean horizon, the sun low and golden, clouds lit pink and amber, gentle waves catching warm light, slow camera drift, cinematic"
Things this example demonstrates
- Card ↔ motion alternation: LTX clip → Ideogram card → LTX clip → Ideogram card, a pattern interrupt roughly every 13 seconds.
- Caption-safe card design: the scattering card was regenerated once to keep its bottom fifth clear — burned captions occupy y≈1640+, so card content must end above that. (The first version had a tagline there; the caption pill covered it.)
- Pacing QC in practice: built-in speakers pace naturally; voice clones
inherit pace from their reference and often need
voice.maxWpm— see the template README.
The captions/*.json Ideogram caption files are committed alongside the
words_*.json timing files as documentation of how each card was authored.