Overview
InvokeAI is a self-hosted creative tool for generating and editing images with open diffusion models. It runs a local web server with a React-based UI, so you work in your browser while everything stays on your own hardware. The project also serves as the foundation for several commercial products.
It is aimed at artists, designers, and developers who want hands-on control over image generation rather than a hosted API. The Unified Canvas supports inpainting, outpainting, and brush tools, while the node-based workflow editor lets you build and share custom generation pipelines.
Within the image-generation category, InvokeAI sits alongside tools like AUTOMATIC1111 and ComfyUI. It pairs a polished UI with a node graph under the hood, and supports a wide range of models including SD 1.5, SDXL, SD 3.5, and the Flux family.
What it does
- Locally hosted web server with a React UI for generating and refining images
- Unified Canvas with inpainting, outpainting, and brush tools for iterative editing
- Node-based workflow editor to build, save, and share custom generation pipelines
- Board and gallery management with rich image metadata for recalling prompts and settings
- Broad model support: SD 1.5, SD 2.0, SDXL, SD 3.5, CogView 4, and the Flux family
- Built-in model manager, upscaling tools, embeddings, and object segmentation (SAM / SAM2)
Getting started
InvokeAI installs through its official Launcher, which handles setup and updates for you on compatible hardware.
Download the Launcher
Grab the latest Launcher for your platform from the InvokeAI releases page on GitHub. It manages installation and updates.
https://github.com/invoke-ai/launcher/releases/latestInstall and launch Invoke
Run the Launcher and follow the prompts to install Invoke on your machine. When it finishes, it starts the local web server.
Open the web UI
Once Invoke is running, open the locally hosted web UI in your browser to import models and start generating images. See the docs for model setup and tutorials.
Commands and code are distilled from the project's own documentation — always check the official repo for the latest.
When to use it
- Generating and iterating on images locally without sending prompts to a hosted service
- Editing existing artwork, sketches, or photos with inpainting and outpainting on the Unified Canvas
- Building repeatable, shareable generation pipelines with node-based workflows for production work
- Managing and experimenting with multiple open diffusion models from one interface
How InvokeAI compares
InvokeAI alongside other open-source image generation tools AI/TLDR tracks, ranked by GitHub stars.
| Tool | Stars | What it does |
|---|---|---|
| Stable Diffusion web UI (AUTOMATIC1111) | ★ 164k | A browser interface for running Stable Diffusion image generation locally with extensions and fine-grained controls. |
| ComfyUI | ★ 118k | A node-based visual editor for building and running image and video generation pipelines like Stable Diffusion and FLUX locally. |
| Fooocus | ★ 50.4k | A simplified image generation app built on Stable Diffusion that hides technical settings for easy prompting. |
| InvokeAI | ★ 27.5k | Self-hosted web UI and canvas for generating images with open diffusion models |
| Stability-AI generative-models | ★ 27.2k | Stability AI's official code for its Stable Diffusion family of image and video generation models. |
| FLUX | ★ 25.6k | Black Forest Labs' open-weight diffusion models and inference code for generating and editing images from text prompts. |
| Z-Image | ★ 11.6k | Alibaba Tongyi's 6B-parameter open image model that produces photorealistic images quickly on a single GPU. |
| DALLE2-pytorch | ★ 11.3k | An open implementation of DALL-E 2 in PyTorch, with the CLIP encoder, diffusion prior, and cascading decoder you train to generate images from text. |