diffusionstudio/editor

What it solves

Diffusion Studio is a professional video editor designed specifically to be driven by AI agents. It bridges the gap between manual visual editing and programmatic control, allowing agents to create, cut, and refine video content through code and command-line interfaces rather than just a manual GUI.

How it works

The editor treats video projects as code, using SolidJS modules (JSX) as the document source. This creates a bidirectional link: changes made on the visual canvas are written back to the code, and changes to the code are immediately reflected on the canvas.

To enable AI agents to operate the editor, it provides a CLI tool called dapi. This CLI allows agents to "see" and "hear" media through tools like frame grabbing, waveform generation, transcription, and multimodal model queries, and then execute edits on the timeline based on that understanding.

Who it’s for

• AI Developers: Those building agents (using tools like Claude Code, Cursor, or Gemini CLI) that can automate video production.
• Content Creators: Users who want to combine generative AI assets (images, video, voiceovers) with traditional video editing.
• Motion Designers: People creating explainers and promos using a code-first approach.

Highlights

• Bidirectional Editing: Synchronized updates between a JSX-based code editor and a visual video canvas.
• Agent-Ready CLI: A comprehensive set of dapi commands for media probing, transcription, and scene capture for AI agents.
• Code-as-Composition: Video elements, timing, and generative assets are declared in JSX, making them versionable and editable via script.
• Multimodal Integration: Built-in capabilities to generate images, video, and voiceovers directly within the timeline via code.