digitalsamba/claude-code-video-toolkit
Make videos with Claude Code: script, AI voiceover, music, video clips and render, on open-source models that cost cents
About digitalsamba/claude-code-video-toolkit
digitalsamba/claude-code-video-toolkit is an open-source project on GitHub, mainly written in Python. Make videos with Claude Code: script, AI voiceover, music, video clips and render, on open-source models that cost cents It currently holds 2,482 stars and 432 forks with 9 open issues, and was last pushed on 2026-11-08 (repository created 2025-12-09).
Project Overview
Git Homed tracks it on the Audio Trending board and on the AI Audio Trending list.
GitHub Repository Details
README
claude-code-video-toolkit
Tell Claude Code what video you want: it writes the script, generates the voiceover, music, and visuals, and renders the MP4. An AI-native video production workspace: the skills, commands, templates, and tools your AI agent needs to take a video from concept to final render, on open-source models that cost cents (voiceover ~$0.01, AI video clips ~$0.23). With Blender installed (optional), it also makes films with a 3D character who acts from a JSON spec.
See What It Makes
Every frame below was generated in this workspace. Click one to watch it.
![]()
Just a Quick One: the toolkit's trailer, with a 3D character acting from a JSON spec. He walks into his kitchen
with his tea, sits down and opens the laptop; every move is a verb in the spec, checked before Blender renders it.
How it works, and what your machine needs.
![]()
Made with itself: this toolkit in 90 seconds, made with this toolkit. Every terminal line on screen is the real
output of the command that made that part, and the bill is measured ($0.33).
| | |
|:--:|:--:|
|
|
|
| Release-notes video
product UI rebuilt from source,
no screen recording | Design sources
one chart in three design skills,
plug in your own or use ours |
| | | |
|:--:|:--:|:--:|
|
|
|
|
| Product demo
scripted cursor and camera,
AI narrator tile | AI hallucinations explainer
9:16 Short: Ideogram 4 cards,
cloned voice, burned captions | You're On Mute
every video call ever, as a horror trailer;
invented cast, parody-short skill |
More in the showcase table below, and new ones land on the YouTube channel first (subscribe). Made one yourself? Post it in Show and tell.
Quick Start
git clone https://github.com/digitalsamba/claude-code-video-toolkit.git
cd claude-code-video-toolkit
uv sync # Optional: AI voiceover, image gen, music, moviepy examples
claude # Open Claude Code in the toolkit
Python dependencies are managed with uv —uv synccreates.venv/and installs everything from the lockfile in seconds. No uv yet?curl -LsSf https://astral.sh/uv/install.sh | sh(macOS/Linux) orpowershell -c "irm https://astral.sh/uv/install.ps1 | iex"(Windows).
Then in Claude Code:
/setup # Configure cloud GPU, storage, voice (~5 min, mostly free)
/video # Create your first video
That's it. /setup walks you through everything interactively — cloud GPU provider, file transfer, voice config. /video creates a project from a template and guides you through the whole workflow.
What's free: The toolkit leans heavily on open-source AI models — voiceovers (Qwen3-TTS), image generation (FLUX.2), music (ACE-Step), and more. You deploy them to your own cloud GPU account and run them at cost. Cloudflare R2 has a generous free tier (10GB, zero egress), and Modal gives $30/month free compute on the Starter plan — more than enough for a few 5-minute videos a month.
Requirements: Node.js 22+ (24 LTS recommended) and Claude Code. uv recommended for the AI tools (it installs Python 3.10+ for you). FFmpeg optional.
Want to skip setup and just render something?
> cd examples/hello-world && npm install && npm run render
No API keys needed — outputs an MP4 immediately.
---
A Note from the Author (not AI-generated)
I've spent months painstakingly putting this toolkit together and plan to keep iterating on it. AI makes things easier, but hard work still has huge value. Every video I create is a chance for improvement — every skill, template, tool, and workflow here has been refined through that cycle. It would be wonderful if others wanted to get involved with that: use it, refine it, and feed back into the repo via an issue or PR what you learn.
> My own use case is fairly specific: creating sprint review videos for the AI mobile development arm of Digital Samba. But the idea behind this project is a reusable toolkit for using Claude Code to autonomously generate any kind of "explainer" style video — product demos, walkthroughs, presentations, whatever you need. Autonomous video creation is a lofty ideal for such a subjective field, but we can try :)
> What makes this work is that Claude Code is fantastically resourceful and flexible — give it the framing and tooling that this toolkit provides and it will adapt it to create templates and videos based on your prompting. The skills, templates, and tools here are building blocks. Claude Code is the builder. You are the director, editor, and designer.
> If you're getting started, run/setupthen/videoand let Claude Code guide you. Or start with/templateto create a template for your own use case.
> Cloud GPU — I recommend Modal for running the toolkit's AI tools. The Starter plan gives you $30/month free compute, which is more than enough. RunPod is also supported as an alternative. Run /setup to deploy the tools you need.
> My motto: Be brave. Experiment. And please share any videos you create or ideas you have back with the project — it helps me keep improving this toolkit for everyone.
Features
Skills
Claude Code has deep knowledge in:
| Skill | Description | |-------|-------------| | remotion | React-based video framework — compositions, animations, rendering | | elevenlabs | AI audio — text-to-speech, voice cloning, music, sound effects | | ffmpeg | Media processing — format conversion, compression, resizing | | playwright-recording | Browser automation — record demos as video | | video-design | Visual design for video: colour by act, one data accent, type steps, a contact-sheet critique; takes direction from any design skill (design sources) | | qwen-edit | AI image editing — prompting patterns and best practices | | ideogram4 | AI image generation with best-in-class in-image text — title cards, thumbnails, exact brand colors | | acestep | AI music generation — prompts, lyrics, scene presets, video integration | | ltx2 | AI video generation — text-to-video, image-to-video clips, prompting guide | | parody-short | Comedy shorts about a group's own life or a shared topic, in a borrowed genre: in-jokes from the group's own chat and docs, costumed and animated photos or an invented cast, sound placed on the movement, vertical cuts | | blender (optional) | 3D from a JSON shot spec: a rigged cartoon cast acting on cue, a scene that turns out to be on a laptop, an app on a phone or a 3D model on a studio stage; every camera checked before the render; as a film or as clips in any video. Needs Blender and renders in minutes with a GPU. Characters fitted from a generated 3D model are experimental | | moviepy | Python video composition — overlay text on LTX-2/SadTalker output, build.py-style projects | | runpod | Cloud GPU — setup, Docker images, endpoint management, costs |
The always-current catalog of skills, commands, tools, and templates lives in _internal/toolkit-registry.json.
Commands
| Command | Description |
|---------|-------------|
| /setup | First-time setup — cloud GPU, file transfer, voice, prerequisites |
| /video | Video projects — list, resume, or create new |
| /scene-review | Scene-by-scene review in Remotion Studio |
| /design | Set a video's design direction (design.md) from a design source, refine a scene, or critique the render |
| /brand | Brand profiles — list, edit, or create new |
| /template | List available templates or create new ones |
| /skills | List installed skills or create new ones |
| /contribute | Share improvements — issues, PRs, examples |
| /record-demo | Record browser interactions with Playwright |
| /generate-voiceover | Generate AI voiceover from a script |
| /redub | Redub existing video with a different voice |
| /voice-clone | Record, test, and save a cloned voice to a brand |
| /thumbnail | Generate a YouTube thumbnail with Ideogram 4 (hook phrase + brand colours) |
| /publish | Publish a finished video to YouTube (metadata auto-filled from project.json) |
| /versions | Check dependency versions and toolkit updates |
Note: After creating or modifying commands/skills, restart Claude Code to load changes.
Templates
Pre-built video structures in templates/:
- sprint-review — Sprint review videos with demos, stats, and voiceover
- sprint-review-v2 — Composable scene-based sprint review with modular architecture
- product-demo — Marketing videos with dark tech aesthetic, stats, CTA
- concept-explainer-short — 9:16 vertical TikTok/Reels/Shorts explainers (Python/moviepy, no Remotion)
- ds-release-notes — Synthetic release-notes videos: product screens rebuilt from source, scripted cursor and camera, sentence-level cloned narration (ships as the Digital Samba 650 build)
examples/ for finished projects you can learn from (newest first — scroll down to watch the toolkit evolve in reverse):
| Date | Demo | Description |
|------|------|-------------|
| 2026-11-12 | made-with-itself | 93s video about the toolkit, made in the toolkit from one sentence typed into Claude Code: each stage (narrate, score, generate, compose, check) is the real command in the new Terminal component, every line its saved output, then what it made; narrated by Marilyn from the ElevenLabs voice library through the new brands/video-toolkit; the bill is measured ($0.33) and the Check scene shows what qc.py really found in the draft (examples/toolkit-made-itself) |
| 2026-11-09 | design-sources | 46s: one chart designed three ways (the toolkit's own palettes, taste-skill, and a house style of your own), the page turning between them on the narration with PageTurns, then the render checked one frame per scene with contact_sheet.py |
| 2026-11-07 | the-great-return | 98s wildlife documentary about the return to the office, made with the parody-short skill and the same invented cast: mock-Latin lower thirds, a designed narrator voice, LTX clips lengthened with ltx_extend.py instead of frozen, every sound effect placed on its movement with sync_check.py (vertical cut) |
| 2026-11-07 | youre-on-mute | 76s horror trailer about the video call that never ends, from the parody-short skill: an invented cast of four (FLUX portraits animated with LTX through LivingPhoto), a call UI built in Remotion that misbehaves, a stinger after the title (vertical cut) |
| 2026-11-04 | data-chart-stars | 31s chart video of the toolkit's GitHub star history, drawn as SVG by the new LineChart and BarChart components: the line is drawn while Lugh narrates each stretch, and the March spike's band and the monthly bars light on the words that name them, timed with cue() (examples/data-chart-stars) |
| 2026-11-01 | the-agent-handoff | An experiment in agent-first, one-prompt video, kept for the record rather than as a showcase. 57s an agent made unattended from one Discord message: the agent wrote the script and the shot list, video_job.py voiced it with a cloned voice, generated four LTX-2 clips and a diagram card, built and QC'd it, and the Discord bot posted the link when it went live; $0.47, the fourth take after three pipeline fixes (docs/agent-video-jobs.md) |
| 2026-10-28 | ds-release-650 | 87s release-notes video built from the ds-release-notes template: six items across the room and account center, Lugh's cloned voice made livelier with voice_prosody.py, checked with qc.py --render --layout (templates/ds-release-notes) |
| 2026-10-21 | ds-release-649 | Release-notes video with no screen recording: three product UIs rebuilt from their source code, driven by a scripted cursor and camera, narrated by Lugh; re-rendered 2026-11-09 in the DS act palette from video-design (examples/ds-release-649) |
| 2026-10-21 | ds-free-schedule-demo | 26s synthetic product demo of scheduling a call, with a SoulX talking-head narrator tile (examples/ds-free-schedule-demo) |
| 2026-06-09 | hallucinations-short | 3:41 vertical explainer on AI hallucinations — Ideogram 4 cards, LTX-2 b-roll, burned karaoke captions, Qwen3-TTS voice clone (reference sample from Pixabay) |
| 2026-06-09 | sky-blue-short | 52s "Why is the sky blue?" — concept-explainer-short template showcase (examples/sky-blue-short), stock Qwen3-TTS voice |
| 2026-04-08 | q2-townhall-stars | GitHub star history time-lapse with animated chart and deadpan-to-excited commentary |
| 2026-04-08 | q2-townhall-longarm-ad | Super Bowl-style launch ad with dramatic Qwen3-TTS announcer and LTX-2 animated Lugh cameo |
| 2026-03-15 | the-space-between | AI-generated video essay — flux2 avatar, Qwen3-TTS voice, SadTalker animation |
| 2026-02-23 | cortina | Mobile platforms sprint review |
| 2026-01-25 | schlumbergera | Android sprint review video |
| 2026-01-22 | ds-remote-mcp | Remote MCP server demo (the jazz background music is a joke) |
| 2025-12-10 | digital-samba-skill-demo | Product demo showcasing Claude Code skill |
| 2025-12-05 | sprint-review-cho-oyu | iOS sprint review with demos |
Videos other people have made with the toolkit are in Show and tell. Share yours there: a link, the template or tools you used, and what it cost.
Scene Transitions
The toolkit includes a transitions library for scene-to-scene effects:
| Transition | Description |
|------------|-------------|
| glitch() | Digital distortion with RGB shift |
| rgbSplit() | Chromatic aberration effect |
| zoomBlur() | Radial motion blur |
| lightLeak() | Cinematic lens flare |
| clockWipe() | Radial sweep reveal |
| pixelate() | Digital mosaic dissolution |
| checkerboard() | Grid-based reveal (9 patterns) |
Plus official Remotion transitions: slide(), fade(), wipe(), flip()
Preview all transitions:
cd showcase/transitions && npm install && npm run studio
See lib/transitions/README.md for full documentation.
Brand Profiles
Define visual identity in brands/. When you create a project with /video, the brand's colors, fonts, and styling are automatically applied.
brands/my-brand/
├── brand.json # Colors, fonts, typography
├── voice.json # ElevenLabs voice settings
└── assets/ # Logo, backgrounds
Included brands: default, digital-samba
Create your own with /brand.
Project Management System
Video projects are tracked through a multi-session lifecycle:
planning → assets → review → audio → editing → rendering → complete
Each project has a project.json that tracks:
- Scenes — What to show, asset status, visual types
- Audio — Voiceover and music status
- Sessions — Work history across Claude Code sessions
- Phase — Current stage in the workflow
CLAUDE.md per project for instant context when resuming.
See lib/project/README.md for schema details, scene status tracking, and filesystem reconciliation logic.
Python Tools
Audio, video, and image tools in tools/:
# AI voiceover — ElevenLabs or self-hosted Qwen3-TTS (9 voices + cloning)
uv run tools/voiceover.py --provider qwen3 --speaker Ryan --scene-dir public/audio/scenes --json
AI music (ACE-Step — free cloud API)
uv run tools/music_gen.py --preset corporate-bg --duration 120 --output music.mp3
AI image generation (FLUX.2) and editing (Qwen-Image)
uv run tools/flux2.py --preset title-bg --brand digital-samba --cloud modal
uv run tools/image_edit.py --input photo.jpg --prompt "Add sunglasses" --cloud modal
AI video generation (LTX-2.3 — text-to-video, image-to-video)
uv run tools/ltx2.py --prompt "A sunset over the ocean, cinematic" --cloud modal
Talking head from a portrait + audio (SoulX-FlashHead)
uv run tools/soulx.py --image portrait.png --audio voiceover.mp3 --output talking.mp4
All tools — sound effects, redub, upscaling, watermark removal, NotebookLM rebranding, in-image text…
# Generate voiceover (ElevenLabs)
uv run tools/voiceover.py --script script.md --output voiceover.mp3
Generate voiceover (Qwen3-TTS — self-hosted, cheaper alternative)
uv run tools/voiceover.py --provider qwen3 --speaker Ryan --scene-dir public/audio/scenes --json
uv run tools/qwen3_tts.py --text "Hello world" --tone warm --output hello.mp3
Generate background music (ElevenLabs)
uv run tools/music.py --prompt "Upbeat corporate" --duration 120 --output music.mp3
Generate background music (ACE-Step — free cloud API, XL Turbo 4B model)
uv run tools/music_gen.py --preset corporate-bg --duration 120 --output music.mp3
uv run tools/music_gen.py --prompt "Dramatic cinematic" --duration 30 --bpm 90 --key "D Minor" --output reveal.mp3
uv run tools/music_gen.py --prompt "Upbeat indie rock" --duration 60 --variations 4 --output intro.mp3
Generate sound effects
uv run tools/sfx.py --preset whoosh --output sfx.mp3
Find real recorded sound effects (Creative Commons, free, no API key)
uv run tools/sfx_search.py --query "whoosh" --output sfx.mp3
Redub video with different voice
uv run tools/redub.py --input video.mp4 --voice-id VOICE_ID --output dubbed.mp4
Add background music to existing video
uv run tools/addmusic.py --input video.mp4 --prompt "Subtle ambient" --output output.mp4
Rebrand NotebookLM videos (trim outro, add your logo/URL)
uv run tools/notebooklm_brand.py --input video.mp4 --logo logo.png --url "mysite.com" --output branded.mp4
AI image editing (style transfer, backgrounds, custom prompts)
uv run tools/image_edit.py --input photo.jpg --style cyberpunk --cloud modal
uv run tools/image_edit.py --input photo.jpg --prompt "Add sunglasses" --cloud modal
AI image upscaling (2x/4x)
uv run tools/upscale.py --input photo.jpg --output photo_4x.png --cloud modal
Remove watermarks (requires cloud GPU)
uv run tools/dewatermark.py --input video.mp4 --preset sora --output clean.mp4 --cloud modal
Locate watermark coordinates
uv run tools/locate_watermark.py --input video.mp4 --grid --output-dir ./review/
Generate talking head video from image + audio (SoulX-FlashHead)
uv run tools/soulx.py --image portrait.png --audio voiceover.mp3 --output talking.mp4
AI image generation (FLUX.2 klein 4B; klein-9B is a non-commercial Modal opt-in)
uv run tools/flux2.py --prompt "A sunset over mountains" --cloud modal
uv run tools/flux2.py --preset title-bg --brand digital-samba --cloud modal
uv run tools/flux2.py --list-presets
AI video generation (LTX-2.3 22B — text-to-video + image-to-video)
uv run tools/ltx2.py --prompt "A sunset over the ocean, cinematic" --cloud modal
uv run tools/ltx2.py --prompt "Gentle camera drift" --input photo.jpg --cloud modal
3D shots from a JSON spec (optional; needs Blender 4.2+, and a GPU to render in minutes: see docs/blender.md)
uv run tools/blender_render.py doctor # is Blender here, and how fast is it?
uv run tools/blender_render.py check examples/blender-disk-panic/spec.json # cameras verified, contact sheet, ~5 s
uv run tools/blender_render.py render spec.json --clips --alpha --to projects/my-video/public/blender
Publish a finished render to YouTube (OAuth 2.0 + Data API v3) — contributed by @dascope (#29)
uv run tools/youtube_upload.py --auth # one-time browser login
uv run tools/youtube_upload.py --video out/video.mp4 --title "My video" --privacy private --json-out
Tool Categories:
| Type | Tools | Purpose | |------|-------|---------| | Project | voiceover, music, music_gen, sfx, sfx_search | Used during video creation workflow | | Utility | redub, addmusic,