# Best open-source AI repos for AI video editing and generation (2026)

Source: https://dedupedai.com/best-ai-repos-for/ai-video-editing/ (Deduped, updated 2026-10-26). Markdown twin of the page; JSON: https://dedupedai.com/best-ai-repos-for/ai-video-editing/index.json

## What is the best open-source repo for editing or making videos with AI?

The best open-source AI repos for AI video editing and generation right now are Video Use, OpenMontage and Hyperframes. To edit raw footage by chatting with a coding agent, use video-use; to have an agent plan, generate and render whole videos, use OpenMontage; to turn HTML and animation into MP4, use HyperFrames; Concat is a full local editor.

## Comparison

| Repo | What it does | Best for | Stars | Language |
|---|---|---|---|---|
| [Video Use](https://dedupedai.com/use-cases/browser-use-video-use/) (browser-use/video-use) | video-use lets you edit videos by chatting with Claude Code or another coding agent. | Creators, founders and marketers who record talking heads, tutorials or interviews and want edited videos without a timeline editor. | 29,906 | Python |
| [OpenMontage](https://dedupedai.com/use-cases/calesthio-openmontage/) (calesthio/OpenMontage) | OpenMontage is an open-source system that lets an AI coding assistant plan, generate, edit, and render videos from plain-language requests through 12 production pipelines. | Creators, marketers, and developers who want an AI coding assistant to produce finished videos from a written request. | 68,946 | Python |
| [Hyperframes](https://dedupedai.com/use-cases/heygen-com-hyperframes/) (heygen-com/hyperframes) | HyperFrames is an open-source framework that turns HTML, CSS, media, and seekable animations into deterministic MP4 videos, usable through a CLI or AI coding agent skills. | Marketers, developers, and content creators who want to produce short videos from web content by working with an AI coding agent or the local CLI. | 69,674 | TypeScript |
| [Concat](https://dedupedai.com/use-cases/jub0t-concat/) (jub0t/Concat) | Concat is a free, open-source video editor and CapCut alternative for macOS, Windows, Linux and Android. | Creators making TikToks, Reels, Shorts, YouTube videos and podcast clips who want CapCut features without watermarks, accounts or uploads, and developers who want agents to cut video. | 11,760 | Rust |
| [Open Edit](https://dedupedai.com/use-cases/veedstudio-open-edit/) (veedstudio/open-edit) | Open Edit is an open-source command-line tool that lets people and AI agents create and edit videos, including subtitles, motion graphics, slides, and renders, powered by VEED. | Video editors, marketers, and developers who want an agent or script to caption, cut, render, and finish short videos from the terminal. | 1,694 | TypeScript |
| [Video Shotcraft](https://dedupedai.com/use-cases/vincentwei1021-video-shotcraft/) (Vincentwei1021/video-shotcraft) | video-shotcraft is a skill for Claude Code and Codex that makes product promo videos. | Founders and developers who want a launch film for their product without a video editor. | 13,473 | TypeScript |
| [Motion Graphics Skills](https://dedupedai.com/use-cases/charlie947-motion-graphics-skills/) (charlie947/motion-graphics-skills) | A free set of 15 Claude Code skills that make launch videos, animated charts, reels and other motion graphics by writing the animation as code, with no After Effects or video generator. | Marketers, course creators and newsletter writers who want on-brand launch videos, charts and reels made in Claude Code. | 411 | Shell |

## When to pick which

### Video Use (browser-use/video-use)

video-use lets you edit videos by chatting with Claude Code or another coding agent. Drop raw footage in a folder and get final.mp4 back: it cuts filler words and dead space, auto color grades, adds audio fades at every cut, burns subtitles, generates animation overlays and self-evaluates the render at every cut boundary. The agent reads an ElevenLabs word-level transcript instead of watching frames.

Best for: Creators, founders and marketers who record talking heads, tutorials or interviews and want edited videos without a timeline editor.

Needs: Claude Code, Codex or another agent with shell access; ffmpeg (required) and optionally yt-dlp; Python with uv or pip; An ElevenLabs API key for transcription.

Start here: [Turn a folder of raw takes into a finished launch video](https://dedupedai.com/use-cases/browser-use-video-use/edit-raw-takes-into-a-launch-video/) (Beginner, 30 minutes)

1. Paste the README setup prompt into Claude Code or Codex so it installs the repo, ffmpeg and the skill
2. Give it your ElevenLabs API key when it asks
3. Run cd /path/to/your/videos and start your agent
4. Say edit these into a launch video and approve the proposed strategy

All 3 use cases: https://dedupedai.com/use-cases/browser-use-video-use/ | GitHub: https://github.com/browser-use/video-use

### OpenMontage (calesthio/OpenMontage)

OpenMontage is an open-source system that lets an AI coding assistant plan, generate, edit, and render videos from plain-language requests through 12 production pipelines.

Best for: Creators, marketers, and developers who want an AI coding assistant to produce finished videos from a written request.

Needs: Python 3.10+; FFmpeg installed and on the PATH; Node.js 18+; An AI coding assistant such as Claude Code, Cursor, Copilot, Windsurf, or Codex.

Start here: [Build a real-footage documentary montage for free](https://dedupedai.com/use-cases/calesthio-openmontage/build-real-footage-documentary-montage/) (Intermediate, 45 minutes)

1. Install prerequisites: Python 3.10+, FFmpeg, and Node.js 18+.
2. make setup
3. Open the OpenMontage folder in your AI coding assistant.
4. Send the request: "Make a 75-second documentary montage about city life in the rain. Use real footage only, no narration, elegiac tone, with music."

All 5 use cases: https://dedupedai.com/use-cases/calesthio-openmontage/ | GitHub: https://github.com/calesthio/OpenMontage

### Hyperframes (heygen-com/hyperframes)

HyperFrames is an open-source framework that turns HTML, CSS, media, and seekable animations into deterministic MP4 videos, usable through a CLI or AI coding agent skills.

Best for: Marketers, developers, and content creators who want to produce short videos from web content by working with an AI coding agent or the local CLI.

Needs: An AI coding agent that supports skills, such as Claude Code, Codex, Cursor, or Gemini CLI; Network access to install the plugin or skills from the heygen-com/hyperframes repository; For the PR use case: the gh CLI, which the README says the pr-to-video skill uses to read pull requests.

Start here: [Make a product launch video from your website with an AI agent](https://dedupedai.com/use-cases/heygen-com-hyperframes/product-launch-video-from-url/) (Beginner, 30 minutes)

1. In Claude Code: claude plugin marketplace add heygen-com/hyperframes then claude plugin install hyperframes@hyperframes
2. For Codex, Cursor and other agents run npx hyperframes skills update once
3. Load the router (/hyperframes:hyperframes in Claude Code) and ask for a product launch video from your website URL
4. Preview with npx hyperframes preview and ask for changes in plain words

All 5 use cases: https://dedupedai.com/use-cases/heygen-com-hyperframes/ | GitHub: https://github.com/heygen-com/hyperframes

### Concat (jub0t/Concat)

Concat is a free, open-source video editor and CapCut alternative for macOS, Windows, Linux and Android. It runs locally on a native Rust engine with a GPU compositor and covers auto-captions with local Whisper, text-to-speech and voice cloning, background removal, keyframes, 100 effects and 4K export, with no watermark, account or upload. A built-in MCP server (concat --mcp) and a JSON-RPC API let AI agents and scripts edit video too.

Best for: Creators making TikToks, Reels, Shorts, YouTube videos and podcast clips who want CapCut features without watermarks, accounts or uploads, and developers who want agents to cut video.

Needs: macOS, Windows, Linux or Android (iOS sideloaded); AI models for captions, voices and cutout download once from Settings, then work offline; To build from source: Rust 1.93+, FFmpeg 7+ development libraries, cmake and a C++ compiler.

Start here: [Add styled auto-captions to Shorts and Reels, fully offline](https://dedupedai.com/use-cases/jub0t-concat/caption-short-videos-offline/) (Beginner, 20 minutes)

1. Download Concat from the website or GitHub Releases for your platform
2. On macOS, if an unsigned build will not open, run xattr -dr com.apple.quarantine /Applications/Concat.app
3. Download a caption model size once from Settings
4. Drop in your footage and generate auto-captions; style them with title presets

All 3 use cases: https://dedupedai.com/use-cases/jub0t-concat/ | GitHub: https://github.com/jub0t/Concat

### Open Edit (veedstudio/open-edit)

Open Edit is an open-source command-line tool that lets people and AI agents create and edit videos, including subtitles, motion graphics, slides, and renders, powered by VEED.

Best for: Video editors, marketers, and developers who want an agent or script to caption, cut, render, and finish short videos from the terminal.

Needs: Node.js, so that npx can run @veedstudio/openedit-cli; A workspace folder for the project; FFmpeg, which install-ffmpeg can find or download; A VEED login for hosted transcription and the VEED editor handoff.

Start here: [Caption a video locally with WhisperX at no cost](https://dedupedai.com/use-cases/veedstudio-open-edit/caption-video-locally-with-whisperx/) (Intermediate, 45 minutes)

1. Run npx @veedstudio/openedit-cli init in an empty folder to set up the workspace.
2. Run npx @veedstudio/openedit-cli install-whisperx to install the local provider.
3. Run npx @veedstudio/openedit-cli transcribe --provider whisperx video.mp4 to transcribe the video.
4. Open the runs folder and check the transcript.json file for word timings.

All 5 use cases: https://dedupedai.com/use-cases/veedstudio-open-edit/ | GitHub: https://github.com/veedstudio/open-edit

### Video Shotcraft (Vincentwei1021/video-shotcraft)

video-shotcraft is a skill for Claude Code and Codex that makes product promo videos. The agent captures your real pages, picks shots from a library of tuned motion recipes, storyboards the film, animates it with Remotion, adds sound design and hands you an MP4 you can keep editing. It is not a video generator: every frame is drawn by code from your own screenshots and copy.

Best for: Founders and developers who want a launch film for their product without a video editor.

Needs: Claude Code or Codex; Node for the Remotion project; Your own product pages to capture.

Start here: [Install the Video Shotcraft skill](https://dedupedai.com/use-cases/vincentwei1021-video-shotcraft/install-video-shotcraft/) (Beginner, 10 minutes)

1. Run npx skills add Vincentwei1021/video-shotcraft
2. Or tell your agent: Install this skill for me: https://github.com/Vincentwei1021/video-shotcraft
3. Or by hand run git clone https://github.com/Vincentwei1021/video-shotcraft.git and cd video-shotcraft
4. Link it with ln -s "$(pwd)" ~/.claude/skills/video-shotcraft for Claude Code or the ~/.codex/skills path for Codex

All 3 use cases: https://dedupedai.com/use-cases/vincentwei1021-video-shotcraft/ | GitHub: https://github.com/Vincentwei1021/video-shotcraft

### Motion Graphics Skills (charlie947/motion-graphics-skills)

A free set of 15 Claude Code skills that make launch videos, animated charts, reels and other motion graphics by writing the animation as code, with no After Effects or video generator.

Best for: Marketers, course creators and newsletter writers who want on-brand launch videos, charts and reels made in Claude Code.

Needs: Claude Code, with Opus 5.5 selected via /model; Install the skills with npx skills add charlie947/motion-graphics-skills; Run brand-intake first so brand.md and MOTION.md exist; For MP4 export: HyperFrames (npx skills add heygen-com/hyperframes) or ffmpeg plus Chrome.

Start here: [Make a 30 to 45 second launch video for a product or cohort](https://dedupedai.com/use-cases/charlie947-motion-graphics-skills/launch-video-for-new-cohort/) (Intermediate, 60 minutes)

1. Run npx skills add charlie947/motion-graphics-skills
2. Open Claude Code, pick Opus 5.5 with /model, and run brand-intake first
3. Say: "Make a launch video for my new cohort." Replace the example with your own offer details
4. Check the draft against your approved facts list and ask for edits such as "Slow down the second scene."

All 5 use cases: https://dedupedai.com/use-cases/charlie947-motion-graphics-skills/ | GitHub: https://github.com/charlie947/motion-graphics-skills

## Questions

### What is the best open-source repo for editing or making videos with AI?

The best open-source AI repos for AI video editing and generation right now are Video Use, OpenMontage and Hyperframes. To edit raw footage by chatting with a coding agent, use video-use; to have an agent plan, generate and render whole videos, use OpenMontage; to turn HTML and animation into MP4, use HyperFrames; Concat is a full local editor.

### Which repo for AI video editing and generation is easiest to start with?

Video Use: its starter use case "Turn a folder of raw takes into a finished launch video" is rated Beginner and takes about 30 minutes.

### How do I get started with Video Use?

Paste the README setup prompt into Claude Code or Codex so it installs the repo, ffmpeg and the skill; Give it your ElevenLabs API key when it asks; Run cd /path/to/your/videos and start your agent. Full steps and a copy-paste agent prompt: https://dedupedai.com/use-cases/browser-use-video-use/edit-raw-takes-into-a-launch-video/

### Are these AI video editing and generation repos free?

They are open-source projects on GitHub (check each license; some, like fair-code projects, limit commercial use). Some also sell a hosted version or need a paid model API key; the requirements for each repo are listed on this page.

### How is this list of AI video editing and generation repos chosen?

Deduped tracks the hottest AI repos on GitHub and writes step-by-step use cases from each README. For this page we picked the repos that do this job, ordered by fit and then GitHub stars. Repos we only inferred from a short README, or that we do not index, are left out. The page is rebuilt with the site; stars are as of the last build in 2026.

Third-party repos we don't own or vouch for. AI-written use cases can be wrong: review everything and run at your own risk.