Invoke is a professional creative engine designed to generate and manipulate visual media using Stable Diffusion and related AI models. The project serves as both a standalone application for artists and enthusiasts and as the foundation for multiple commercial products. Built primarily in TypeScript, it provides a web-based interface that runs on locally hosted servers, allowing users to work with AI image generation without relying on external cloud services.
The repository demonstrates significant community engagement and active maintenance. GitGenius tracking shows 1469 issues and pull requests with a median response latency of 0.0 hours, indicating rapid triage and attention to incoming requests. The most active contributor, psychedelicious, has logged 1375 events, while other key contributors like Pfannkuchensack and hipsterusername have maintained consistent involvement. Bug reports comprise the largest category of tracked issues at 848 items, followed by enhancement requests at 563, suggesting the project balances stability improvements with feature development. The project's contributor base overlaps with major technology repositories including Microsoft's VSCode and TypeScript implementations, as well as the Rust language repository, indicating involvement from developers working across multiple significant open-source ecosystems.
The core functionality centers on a unified canvas implementation that integrates all major generation capabilities including text-to-image, image-to-image, inpainting, and outpainting operations. The canvas supports brush tools and allows artists to use AI as a collaborative creative tool for augmenting AI-generated imagery, sketches, photography, and renders. Beyond basic generation, Invoke includes a node-based workflow system that enables users to create customizable generation pipelines, which can be developed and shared within the community for specific production use-cases.
The platform supports an extensive array of models spanning multiple generations and architectures. These include Stable Diffusion variants at versions 1.5, 2.0, and SDXL, as well as newer models like SD 3.5 Medium and Large. The system also integrates Flux model variants including Flux.1 Dev, Schnell, Kontext, Krea, and Redux, alongside specialized models like CogView 4, Z-Image Turbo, Anima, Qwen Image, and others. This broad model support allows users to select the most appropriate architecture for their specific creative needs.
Additional features include a board and gallery management system for organizing, storing, and remixing generated content, with rich metadata preservation enabling easy recall of prompts and settings. The platform supports multiple model formats including ckpt, diffusers, and some gguf files. Supplementary tools encompass upscaling capabilities, an embedding manager, a model manager, and object segmentation and selection models using SAM and SAM2 technology.
The project operates under a commercially-friendly license and is available for download and installation on compatible hardware. Development is sustained through community sponsorship, with the README explicitly encouraging financial support to fund maintenance, new features, and community support. The project maintains active documentation, contribution guidelines, and community channels including Discord for user support and GitHub discussions for development coordination.
