Summary AI CanvasPro is a native web-based (HTML/CSS/JS) multimodal AI editor. It enables users to visually compose and chain various AI capabilities—text, image, video, audio generation—on an infinite canvas by connecting AI nodes, offering a flexible workflow for creative and technical users.

How It Works The core is a node-based architecture implemented as a pure web application. Users interact with an infinite, zoomable canvas where AI functionalities are represented as nodes. Connecting these nodes visually allows data and results to flow between them, constructing complex AI workflows through a graphical interface and facilitating experimentation with different AI model combinations.

Quick Start & Requirements Installation is available via source code for developers or a one-click Windows bundle for general users.

• Source: Requires Git, FFmpeg, Python 3.12+. Clone, install dependencies (pip install -r requirements.txt), and run python server.py. Access via http://localhost:8777.
• Windows Bundle: Download, extract to a non-Chinese path, and run the .bat file (as administrator for first launch). Access via http://localhost:8777. Official download links for Git, FFmpeg, and Python are provided.

Highlighted Details

• Infinite Canvas: Features zoom, pan, mini-map, grid alignment, multi-canvas support, and fit-to-canvas.
• AI Node Types: Supports image (e.g., Banana Pro, GRSAI), text (e.g., Gemini, GPT, OpenAI compatible), video (text/image-to-video), and audio (TTS) generation.
• Project Management: Allows multi-project switching, local JSON saving for migration, and auto-caching.
• API Integration: Configurable API keys support Jiemeng, RunningHub, APImart, PPIO, GRSAI, and OpenAI-compatible endpoints.
• Node Referencing: Enables dynamic embedding of other nodes' outputs into prompts using @ syntax.

Maintenance & Community Maintained by Ashuo (阿硕), with contact via WeChat and Bilibili. Feedback via Feishu link. User documentation in 使用说明.md.

Licensing & Compatibility Released under the MIT License, permitting broad usage, modification, and distribution, including for commercial purposes and closed-source projects.

Limitations & Caveats Jiemeng API access requires advanced membership. Project is actively developed. Primary documentation is in Chinese.