Purpose
InvokeAI is an open-source AI image generation engine for creative professionals, designed to provide production-grade image generation and editing tools for professionals and enthusiasts. It runs in a self-hosted manner, allowing users to complete the entire workflow from generation to post-editing on their own hardware.
Main Capabilities
- Layered canvas editing: Provides a true layered canvas for drawing, painting, sketching, and editing works, with each layer independently operable, facilitating fine composition of complex scenes.
- Node workflows: Build complex, reproducible generation pipelines through a fully visual graph backend, and expose custom UI parameters, allowing parameters to be shared and updated without diving deep into the graph structure.
- Mainstream model support: Out-of-the-box support for the latest base models such as Flux, SDXL, and SD 1.5, with an intuitive model manager for managing checkpoints, LoRA, textual inversion, and ControlNet.
- Deep ControlNet integration: Use depth maps, edges, poses, and more to guide generation and achieve precise composition control.
- Rapid iteration: Supports batch generation, prompt wildcards, and high-resolution upscaling, allowing concepts to be iterated quickly without leaving the application.
- Local and private: Runs fully locally and self-hosted, with data, prompts, and creative outputs entirely owned by the user, free from cloud service restrictions.
How to Use
Invoke Launcher is officially provided for quick installation on Windows, macOS, and Linux. Advanced users can use Docker or manual Python installation. For specific installation commands and steps, please refer to the official documentation.
Use Cases
Suitable for creators and teams that need to run AI image generation locally or on a VPS, value privacy and data control, and want to use professional-grade canvas editing and node workflows.
Limitations
- The Windows version requires Windows 10 or later, as well as an NVIDIA or AMD GPU;
- The macOS version requires Apple Silicon (M series) and is not compatible with Intel;
- The Linux version requires an NVIDIA or AMD GPU and is compatible with most distributions;
- The original official hosting platform Invoke.ai has been shut down, and the project is now maintained by the community;
- The project uses the Apache 2.0 license.
Community and Training
Invoke is continuously developed by an active open-source community, and users can participate in discussions, report issues, or request features. Invoke Training provides a dedicated UI for textual inversion and LoRA training; the official project also provides an FAQ and contribution guidelines.
