# Diffusion Studio

> Open-source video editor combining an infinite canvas with AI agent workflows

Diffusion Studio is an open-source video editor featuring an infinite spatial canvas, multitrack timeline, and built-in AI media generation. It connects with coding agents via MCP and CLI to let tools like Claude Code or Cursor analyze footage, cut clips, and generate audio or visuals directly in the project codebase.

Website: https://diffusion.studio/

## What it is for

- Generate and replace visual elements in video scenes using AI image and video models
- Delegate video editing tasks like clipping, transcription, and sound design to coding agents
- Organize complex multimedia campaigns and multi-format scenes on an infinite canvas
- Export high-resolution 4K video compositions with keyframe easing and effects

## Features

- Infinite canvas paired with a multitrack timeline editor
- Coding agent integration via Model Context Protocol (MCP) and CLI
- Built-in generation for images, video, sound effects, and voiceover
- Property inspector with bezier curve animation and layout control
- Support for HTML, WebGL, 3D elements, and shaders in compositions
- Non-linear multitrack video and audio editing with 4K export support
- Open-source codebase licensed under MPL-2.0

## Pricing

Freemium. Core editor, CLI, and exports are free and open source; paid tiers purchase credits for hosted AI model generation

## Details

- Type: Web app
- Category: [AI](https://xiry.app/topics/ai.md)
- Open source: yes
- Platforms: Web, macOS, Windows, Command line
- For: Video creators, motion designers, and developers building AI-assisted video pipelines
- Alternative to: CapCut, Runway, Adobe Premiere Pro, Descript
- Tags: video editor, ai video, infinite canvas, open source, agentic ai, mcp, motion design

## Why it was saved

- [@konstipaulus, 29 June 2025](https://x.com/konstipaulus/status/1939349264263430428): Described as 'Figma meets Sora': a visual canvas where you drag assets, prompt an agent to generate/replace elements (e.g. swap a Minecraft block into an ASMR photo with FLUX.1 CONTEXT-MAXM, cut videos with KLING-2.1-Standard, add sound with MMAUDIO-V2), approve results, and edit on a Timeline with export.

## Compare

- [Alternatives to CapCut](https://xiry.app/alternatives/capcut.md)
- [Alternatives to Descript](https://xiry.app/alternatives/descript.md)
- [Alternatives to Runway](https://xiry.app/alternatives/runway.md)
- [Alternatives to Adobe Premiere Pro](https://xiry.app/alternatives/adobe-premiere-pro.md)

## Related

- [Agent Reach](https://xiry.app/resources/agent-reach): Open-source capability layer giving AI agents access to browse and scrape platforms (GitHub repo, Open source)
- [OpenManus](https://xiry.app/resources/openmanus): Open-source general AI agent framework replicating Manus without invite codes (GitHub repo, Open source)
- [VideoSOS](https://xiry.app/resources/videosos): Open-source browser-based AI video editor powered by 100+ models (GitHub repo, Open source)
- [Captions](https://xiry.app/resources/captions): AI-powered video editor that automates subtitles, cuts, and visual polish. (Mobile app, Freemium)
- [Submagic](https://xiry.app/resources/submagic): AI video editor for automated captions, B-roll, and short-form clips (Web app, Freemium)
- [invideo](https://xiry.app/resources/invideo): AI-assisted multitrack video editor and text-to-video generator for creators and teams. (Web app, Freemium)

Updated 10 October 2026. Page: https://xiry.app/resources/diffusion-studio

Make your own library: Xiry takes your saved reels, tweets, Reddit bookmarks and Watch Later videos, makes them searchable and pulls the useful stuff out, like this library. It's invite-only for now; join the waitlist at https://xiry.app/#waitlist.