DonkeyCut/Donkey
Open source CapCut alternative. Contribute to DonkeyCut/Donkey development by creating an account on GitHub.
61 stars · 2 forks
README
Donkey Cut A free, open source CapCut alternative. Edit with chat. Generate video, images, voiceovers, and music.
Donkey Cut is a free video editor that runs in your browser. You can save projects in the cloud and open them anywhere. You can also keep projects on your computer so your video files stay local.
Donkey Cut Donkey Cut is a simple video editor with a multi-track timeline, captions, music, effects, and an AI assistant that can help you edit.
The editor with "The Railway Mystery" open. Generated shots in the side panel, with clips and music on the timeline.
Your computer or the cloud You choose where each project lives. Cloud projects work anywhere. You do not need to install anything. Upload your media from the browser and export your video right in the tab. You can also share a project with a read-only link. Local projects keep your files on your computer. Nothing needs to be uploaded. Both types of projects appear together on the same home screen. The AI assistant works with both. You can generate images, video, voiceovers, and music. Generated files are added to your project just like any other media. For Mac users Install the Donkey Cut companion app to connect the editor to your Mac.
Projects are saved in ~/Movies, so you can see and manage them directly in Finder. Captions and transcription run locally on your Mac. Connect the AI assistant to your existing Codex or Claude Code login. Record your screen directly onto the timeline.
When the app is running, cloud projects can also use your Mac for transcription. Generate what you can't shoot Describe a shot in chat and keep iterating until you get what you want. These are the two example projects from the landing page. The prompts are included below. The Railway Mystery is a 1920s comic-style chase. It uses three generated shots with a brass-and-strings score.
Franco-Belgian comic style, early-1900s animation with film grain: a steam train races a cliffside railway through a mountain canyon; a cloaked figure rides the carriage roof; a boy on a bicycle gives chase
Canyon run On the roof Bicycle chase
City poster series uses matching hand-painted travel posters. The posters are animated into 4-second clips and edited together with captions and a waltz.
Hand-painted travel poster, PARIS — woman in a trench coat crossing the street, Eiffel Tower behind, café awnings, 'Live the romance' in red script
Paris — Live the romance New York — Rise above the city
The poster series in the editor. Both posters are animated into clips with captions and a waltz.
How it works The editor looks and works the same everywhere. Each project decides where its files are stored and where processing happens. browser (the editor) ├─ local project ───▶ the browser's own storage, on your computer │ ├─ cloud project ───▶ hosted APIs · Postgres · R2 storage │ └─ with the Mac app ▶ Cut engine on 127.0.0.1 local disk (~/Movies) · bundled ffmpeg · on-device speech · your claude/codex logins
The local engine only runs on your Mac. On a hosted deployment, these routes return 404 before any handler runs. For the full architecture, see docs/guides/cut/README.md. Pricing The editor is free. We only charge for cloud storage and AI content generation.
Repository layout
Path What's there
apps/Donkey The macOS companion app. It runs the Cut engine and handles screen recording.
site The Next.js site, Cut editor, engine, cloud backend, and hosted API routes.
docs Product documentation and engineering guides.
Build and run Run the editor locally: cd site npm install npm run db:generate npm run dev Then open http://localhost:3000/cut. Run the macOS app in development: cd apps/Donkey swift run Donkey Build the packaged app and installer disk image: ./scripts/package-donkey-app.sh open dist/Donkey.app The site uses Supabase Postgres through Prisma. Keep local credentials in .env. Never commit them. Documentation docs/README.md is the source of truth for supported behavior. Good places to start:
Cut for the editor, local engine, and cloud projects. Install Donkey Locally for building the app bundle.
License Apache 2.0. See LICENSE.