Kimu - The open-source AI video editor by the trykimu team

• Install dependencies: prepare Node.js 20+, Python 3.12+, pnpm, and Docker.
• Clone the repository: run git clone https://github.com/trykimu/videoeditor.git and enter the project directory.
• Start services: run docker compose up -d to automatically launch the frontend, backend, database, and rendering services.
• Access the app: open in your browser http://localhost:8080 to use the full functionality.
• Connect a large model: in .env , fill in GEMINI_API_KEY to enable AI editing, or set LLM_PROVIDER=ollama to use a local model for zero cost.
• Log in to your account: complete registration and login via Google OAuth.
• Upload assets: import video/audio/images into the media library, and the system will automatically categorize and organize them.
• Issue commands: open the "Ask Kimu" chat box and describe your needs in natural language (e.g., "cut the first 10 seconds and add a fade-in").
• Real-time preview: AI-generated results appear immediately on the timeline; dragging and zooming are visible in real time.
• Manual fine-tuning: Refine details on unlimited tracks; all edits are non-destructive and traceable.
• Render final cut: Choose MP4/4K and other presets for one-click export, or generate a time-limited share link to send to your team for review.

Kimu's core advantages

• Natural language editing: Replace tedious clicking with chat; AI directly generates cuts, timing, and layout, with an extremely low barrier to entry.
• Completely open source and free: Licensed under an open source license, core features are zero-cost and not cut down, so individuals and small teams can use it with confidence.
• Controllable local deployment: One-click self-hosting with Docker; your media and data remain fully yours, meeting enterprise data compliance requirements.
• Real-time WYSIWYG: Low-latency rendering based on Remotion; every adjustment previews instantly, saying goodbye to render progress bars.
• Video as code: Remotion's declarative architecture makes templates programmable and parameterizable for batch generation, and allows developers to deeply extend the code.
• Multi-user online collaboration: Cross-device cloud sync + cursors/changes visible in real time; teams can review together without sending files back and forth.
• Flexible AI costs: Supports Gemini pay-as-you-go or Ollama running locally for free, so monthly costs can be pressed close to zero.

Kimu's project address

• GitHub repository：https://github.com/trykimu/videoeditor

Comparison of Kimu and similar competitors

Kimu's application scenarios

• Daily short video updates for self-media: AI completes the rough cut, trimming talking-head segments and removing silence; creators only need 90 minutes for fine-tuning, greatly compressing 6 hours of work.
• Rapid Vlog production: After uploading materials, give voice commands to "cut out pauses and add transitions and subtitles," suitable for high-frequency travel and lifestyle bloggers.
• Small team collaborative review: Marketing teams can co-edit in real time, with cursors and changes visible in sync; brand videos no longer require back-and-forth file transfers, and versions can be rolled back.
• Enterprise training and compliance production: Local deployment for industries with data compliance requirements such as finance and healthcare, so sensitive materials never leave the company intranet.