Overview
claude-code-video-toolkit is a video production workspace built for Claude Code. You tell the agent what video you want, and it writes the script, generates the voiceover, music and visuals, and renders the MP4. The repository supplies the pieces the agent needs to do that: skills that teach it Remotion, FFmpeg, ElevenLabs, Playwright recording and several generative models, slash commands such as `/video`, `/scene-review` and `/publish`, ready-made video templates, brand profiles and a set of Python CLI tools.
The generative work leans on open-source models that you deploy to your own cloud GPU account and run at cost: Qwen3-TTS for voiceover and voice cloning, FLUX.2 for images, ACE-Step for music, LTX-2.3 for text-to-video and image-to-video clips, and SoulX-FlashHead or SadTalker for talking heads. Each of these tools runs on Modal (the recommended option) or RunPod via a `--cloud` flag, and the README lists per-run estimates such as about $0.01 for a voiceover and about $0.23 for an LTX-2 clip. ElevenLabs is supported as a paid alternative for voice, music and sound effects.
Composition happens in Remotion, a React-based video framework, with a shared library of components, a theme system and a transitions library (glitch, RGB split, zoom blur, light leak, clock wipe, pixelate, checkerboard, plus Remotion's own slide, fade, wipe and flip). A multi-session project system tracks each video through planning, assets, review, audio, editing, rendering and complete in a `project.json`, reconciles the plan against the files that exist, and writes a per-project `CLAUDE.md` so a later session can resume with context. The author built it for sprint review videos at Digital Samba, but pitches it as a reusable base for any explainer-style video; migration scripts also install its skills and workflows for Codex and Kiro CLI.
What it does
- Claude Code skills for Remotion, ElevenLabs, FFmpeg, Playwright recording, frontend design, Qwen image editing, Ideogram 4, ACE-Step, LTX-2, MoviePy and RunPod
- Slash commands covering the whole workflow: `/setup`, `/video`, `/scene-review`, `/design`, `/brand`, `/generate-voiceover`, `/redub`, `/voice-clone`, `/record-demo` and `/publish` to YouTube
- Python tools for voiceover, music, sound effects, image generation and editing, upscaling, video generation, talking heads, watermark removal and redubbing, with cloud GPU runs on Modal or RunPod
- Templates for sprint reviews, product demos and 9:16 vertical concept explainers, plus brand profiles that apply colours, fonts and voice settings to every project
- Multi-session project tracking with a `project.json` lifecycle and an auto-generated per-project `CLAUDE.md`
- A scene transitions library on top of Remotion's official transitions
Getting started
You need Node.js 18+ and Claude Code; uv is recommended for the Python AI tools and FFmpeg is optional. A hello-world example renders an MP4 without any API keys.
Try a render with no setup
The hello-world example needs only Node.js and outputs an MP4 immediately. `npm run studio` previews it in the browser first.
cd examples/hello-world && npm install && npm run renderClone the toolkit and install the Python tools
`uv sync` creates `.venv/` and installs the locked dependencies for voiceover, image generation, music and the MoviePy examples. Optional extras add Whisper captions or the Modal CLI.
git clone https://github.com/digitalsamba/claude-code-video-toolkit.git
cd claude-code-video-toolkit
uv sync
uv sync --extra modal # optional: Modal CLI for self-hosted cloud GPURun the setup wizard, then make a video
Open Claude Code inside the toolkit. `/setup` walks through cloud GPU provider, file transfer (Cloudflare R2) and voice configuration; `/video` creates a project from a template and brand and guides you through script, assets, audio and render.
claude
# then, inside Claude Code:
/setup
/videoCall the tools directly
Every generator is also a standalone CLI; add `--cloud modal` or `--cloud runpod` to run the GPU-backed ones.
uv run tools/voiceover.py --provider qwen3 --speaker Ryan --scene-dir public/audio/scenes --json
uv run tools/music_gen.py --preset corporate-bg --duration 120 --output music.mp3
uv run tools/ltx2.py --prompt "A sunset over the ocean, cinematic" --cloud modalPreview and render
Inside a project, preview live in Remotion Studio and render the final MP4.
npm run studio
npm run renderCommands and code are distilled from the project's own documentation — always check the official repo for the latest.
When to use it
- Produce recurring sprint review or product demo videos with consistent branding and voiceover
- Turn a topic into a vertical TikTok, Reels or Shorts explainer with generated images, b-roll and burned captions
- Record a browser walkthrough with Playwright and cut it into a narrated demo
- Redub an existing video with a different or cloned voice, or add background music to it
Version history
Every verified update to claude-code-video-toolkit that AI/TLDR tracked, newest first — each links to our coverage and the official changeset.
- 2026-09-08v0.20.1
Fixes NameError crashes in the Codex and Kiro migration scripts and in `dewatermark --setup`.
- 2026-08-31v0.20.0
Makes SoulX-FlashHead the toolkit's talking head generator (`tools/soulx.py`, Modal-only) and adds five Modal/R2 cloud-infrastructure fixes.
How claude-code-video-toolkit compares
claude-code-video-toolkit alongside other open-source video generation tools AI/TLDR tracks, ranked by GitHub stars.
| Tool | Stars | What it does |
|---|---|---|
| MoneyPrinterTurbo | ★ 134k | A self-hosted WebUI, API and CLI that turns one topic or keyword into a finished short video — script, stock or generated footage, voiceover, subtitles and music. |
| OpenMontage | ★ 70.9k | An agentic video production system that gives a coding assistant 12 pipelines and 100+ tools to research, script, generate and render a video end to end. |
| HyperFrames | ★ 70k | HeyGen's HTML-to-video renderer: author a composition in HTML, CSS and JS, preview it in a browser, then render the same input to a deterministic MP4 locally or on AWS Lambda. |
| Open Generative AI | ★ 31.3k | MIT-licensed desktop and web studio for AI image, video and lip-sync generation that runs hosted models through the Muapi.ai API, with optional local image models in the desktop app. |
| Hypit | ★ 24.8k | A markup language and compiler that lets a coding agent rebuild a reference video as an editable workflow — footage, captions, B-roll and effects anchored to words rather than timestamps — and render variants in one command. |
| Toonflow | ★ 19k | Infinite-canvas workbench for AI short-drama production, taking a novel through chapter events, script, storyboards and video nodes. |
| waoowaoo | ★ 14.7k | Self-hosted creative workspace for AI image and video: a canvas holding references and generated assets, a side Assistant that develops the brief, and model-aware controls over your own OpenRouter key. |
| claude-code-video-toolkit | ★ 2.3k | A video production workspace that lets Claude Code script, voice, score and render an MP4 |