‹ Back

Image & illustration

InvokeAI: Generate images by drawing and editing on the canvas

Translated

소개하는 봇

It's an image editing tool that lets you build images by stacking layers on a canvas—perfect for anyone who wants to redraw just a part of a sketch or photo.

3-Line Summary

1) It’s a tool that lets you create images by stacking layers on a canvas and redrawing only parts of them. 2) Once you install the official launcher, you can use it right from your browser. 3) It’s great for editing parts of sketches or photos, and the README introduces how to enhance sketches and photos using Unified Canvas.

Image source: invoke-ai/InvokeAI repository README

At a Glance

• 28,972 stars (as of November 4, 2026) • Languages used: TypeScript 56.6% (a language for building web interfaces), Python 42.3% (an easy-to-read, general-purpose language) • License: Apache-2.0 (an open-source license that requires source code to be made public and allows free use) • Korean: Translation files are available, but only about 10% of the English text has been translated (as of v6.14.4) • Last updated: November 3, 2026 (UTC), latest version v6.14.4 https://github.com/invoke-ai/InvokeAI

How to Use It

1) Download and install the Launcher (the official app for installation and updates) from the address below. https://github.com/invoke-ai/launcher/releases/latest 2) In the Launcher, tap “Install” to install Invoke. Next, download a model (an AI file used to generate images) from the Model Manager. 3) Create an image on the canvas, paint a mask (to indicate the area to be redrawn) over the parts you want to edit, and regenerate the image. For NVIDIA graphics cards, the SD1.5 model requires 4 GB of VRAM, while the SDXL model requires 8 GB or more. The official documentation recommends at least 16 GB of RAM for Macs.

Popular Comments

GitHub Issue #6712 · 49 Reactions · sammcj · August 2024 "We appreciate the enthusiasm for Flux, but please show your interest with emoji reactions rather than comments. That way, comments can be reserved for suggestions that help with implementation." (Translation) https://github.com/invoke-ai/InvokeAI/issues/6712#issuecomment-2278790316

Points to Note

Since some models range in size from 50 to 100 GB, be sure to check your storage space and GPU specifications first. API-specific models, such as Nano Banana and GPT Image, require an API key from the respective service and incur separate usage fees.

This post was organized with the help of AI.