InvokeAI

It provides a feature-rich web interface with capabilities that go well beyond basic prompt-to-image generation including a node-based workflow editor (Unified Canvas), ControlNet integration for pose and structure guidance, IP-Adapter for reference image style transfer, fine-tuned model support, and seamless inpainting and outpainting.

The platform's Unified Canvas allows artists to work with AI generation as part of a broader compositing workflow painting masks, placing reference elements, generating specific regions of a composition, and iterating non-destructively.

InvokeAI supports a wide range of model formats including Stable Diffusion 1.5, SDXL, SD3, and Flux, with model management tools for downloading, organizing, and switching between models and LoRA adapters.

The node editor exposes the full generation pipeline as a visual graph, enabling advanced users to build custom workflows that chain multiple models and processing steps.

InvokeAI is open-source under the Apache 2.0 license and runs locally on consumer hardware with NVIDIA CUDA, AMD ROCm, or Apple Silicon Metal backends.

It is one of the most actively maintained projects in the open-source image generation ecosystem, with regular releases adding support for new architectures and community-contributed nodes.

Professional users choose InvokeAI over cloud-based alternatives for its privacy guarantees (all processing is local), absence of content filters on locally-run models, and the depth of its professional workflow tooling.