AI GitHub
InvokeAI for image generation and canvas editing
Generate and revise images through a canvas and reusable workflows.
Key Takeaways
- The application organizes generation settings and output assets and supports editing operations that constrain where changes appear.
- Choose the supported model and edit operation before bringing in a production asset.
- Use a small canvas edit with a clearly marked region.
- Output depends on checkpoint, conditioning and the selected editing operation.
How it works
The application organizes generation settings and output assets and supports editing operations that constrain where changes appear. A workflow can express steps used repeatedly.
What you need
Choose the supported model and edit operation before bringing in a production asset. Use a simple mask and a baseline generation to check the effect of canvas settings.
What to try first
Use a small canvas edit with a clearly marked region. Compare changes inside and outside the mask, then save the settings that produced the result.
Preserve a copy of the mask and starting image. That helps distinguish effects of the edit operation from randomness when revisiting the output.
Where to look for failures
Output depends on checkpoint, conditioning and the selected editing operation. Keep the original image and masks when a result needs to be recreated.
Related Franklin pages
Start with a small workflow that can be rerun. Keep the seed, mask, model revision and graph together so the visual output can be compared after an edit.
- Study example workflows that show how ComfyUI nodes and models fit together. ComfyUI Examples for inspecting working graphs
- Build reusable generation graphs around models, samplers and conditioning inputs. ComfyUI for node-based image generation
Franklin’s image tests give you examples of outputs to compare across visual models. Franklin’s image tests
Project documentation
Try it for your workflow
Explore the project’s documentation for setup and usage, or save this page to a collection for later.
Read the project documentation ↗