InvokeAI
invoke-ai · Images
GitHub stars
28,327Stats updated
What is InvokeAI?
Invoke brings generation, selective editing, workflow design, model management, and asset organization into one locally hosted creative environment. Artists can work directly on the Unified Canvas, assemble reusable node graphs for repeatable pipelines, and retain prompts, settings, and related outputs in boards and galleries.
InvokeAI is a creative engine for generating visual media with Stable Diffusion models. Its web interface provides a canvas and workflows for working with AI-generated images.
Core capabilities
- Generate and revise images on a unified canvas with masks and brush tools
- Build and share reusable generation pipelines through node-based workflows
- Organize outputs, prompts, settings, and metadata in boards and galleries
A canvas for selective editing
The Unified Canvas joins core generation with inpainting, outpainting, masks, brushes, and other visual controls. This lets creators treat generated imagery, sketches, photography, and renders as editable source material rather than a sequence of disconnected prompt results.
- canvas
- inpainting
- outpainting
- masks
Workflows with asset memory
Node-based workflows turn a generation sequence into a reusable production asset, while boards and galleries keep outputs and metadata close to the process that made them. Model management, upscaling, segmentation, and a locally hosted web service provide the surrounding operational layer. Some listed models are API-only, so a locally hosted interface does not guarantee local processing for every model choice.
- nodes
- workflows
- gallery
- models
Where it fits
Use cases
- 01
Iterate on a visual concept
Generate source material, bring an output onto the Unified Canvas, then use masks, inpainting, outpainting, brush tools, control layers, and model choices to revise selected areas without abandoning the working context.
- 02
Standardize a production pipeline
Combine generation and processing nodes into a reusable workflow so a team can repeat a reviewed sequence, expose meaningful inputs, and share the pipeline without recreating every step manually.
- 03
Manage a library of generated assets
Store outputs in boards and galleries, drag existing images back into supported controls, and use retained prompts and generation metadata to recall, compare, or remix earlier work inside the same workspace.