ddcat-ai/open-ai-canvas

What it solves

It provides an integrated workspace for AI-driven film and short-drama creation, bridging the gap between a text-based brief and final visual assets. It streamlines the creative process by combining scriptwriting, storyboarding, and asset generation into a single workflow.

How it works

The project uses a free-form infinite canvas as its primary interface, where creators can organize projects, nodes, and connections. It integrates a backend (built with Go and Gin) that manages asynchronous task queues for generating text, images, video, and audio. It also supports a local Canvas Agent and MCP tools to connect the web interface with local CLI or Codex apps, allowing for more advanced automation and skill integration.

Who it’s for

It is designed for filmmakers, short-drama creators, and AI artists who need a structured yet flexible environment to manage the entire production pipeline from script to screen.

Highlights

• Infinite Canvas: Supports nodes, connections, zooming, and a minimap for visual project organization.
• Film-Specific Workflow: Dedicated tools for scripts, characters, scenes, style boards, and structured storyboards.
• Multimodal Generation: Integrated support for text, image, video, and audio generation, including advanced video features like first/last frame reference and camera movement.
• Agent Collaboration: Includes a local Canvas Agent and MCP tool support for extending functionality via plugins.
• Asset Management: A built-in material library with resource reference validation and cross-device synchronization.