tigerowo/infinite-canvas

What it solves

Infinite Canvas is an open-source creative workbench designed for the iterative production of images, videos, and audio. It solves the problem of fragmented workflows by consolidating canvas-based orchestration, AI generation tools, reference image editing, and prompt libraries into a single interface, allowing creators to explore visual schemes and iterate on results continuously.

How it works

The project provides a node-based infinite canvas where users can drag, scale, and connect nodes to organize their creative process. It integrates with OpenAI-compatible APIs for image and response generation, supporting features like image-to-image and reference image editing. It also includes a "3D Director's Desk" for arranging characters, models, and environments in a 3D scene to generate camera views that feed back into the canvas as image nodes.

Who it’s for

Visual creators, AI artists, and designers who need a spatial organization tool to manage complex AI-driven multimodal generation workflows and iterate on visual concepts.

Highlights

• AI Agent Integration: An agent that analyzes selected nodes and their upstream dependencies to plan and generate content, inserting results back into the canvas.
• 3D Director's Desk: A specialized 3D environment for managing camera positions, focal length, and aperture, which automatically writes parameters into generation prompts.
• Multimodal Generation: Support for image, video, and audio generation via compatible APIs.
• Panoramic Image Support: Ability to generate or import 2:1 panoramic images to serve as backgrounds for the 3D Director's Desk.
• Comprehensive Prompt Library: A curated collection of hundreds of image prompts organized by case studies.
• Creative Workflows: Support for personal and public templates, variable forms, and AI-created workflows.