InvokeAI invoke-ai/InvokeAI

Local web UI for image generation with a unified canvas for inpainting and outpainting, node workflows and a model manager.

Creator kit + Compare GitHub ↗ Install guide (official) ↗

InvokeAI at a glance

Checked 15 Oct 2026

Invoke runs a local web server and React UI with a Unified Canvas (inpainting, outpainting, brush tools), node workflows, upscaling and a model manager. It supports SD 1.5, SDXL, SD3.5, Flux.1, Flux.2 Klein, Z-Image, Qwen Image and more, plus API-only models (GPT Image, Nano Banana, Wan). Code is Apache-2.0; weights licence depends on the model you install. Published minimums: 4GB VRAM/8GB RAM for SD1.5, 8GB/16GB for SDXL, 10GB/32GB for FLUX.1.

Cite this pageRepoQuarry. “InvokeAI.” Checked 15 Oct 2026. https://repoquarry.dev/projects/invokeai

How much of Midjourney does it replace?

3 of 4 jobs. Generation, canvas inpainting/outpainting and upscaling with open models; style reference not documented in the README.

Text-to-image Partial replacement

Image editing (inpainting and outpainting) Partial replacement

Upscaling Partial replacement

Style and image reference Not covered

For style and image reference, see other Midjourney alternatives. You may also need to build an interface, an API server and hosting yourself.

How much of Adobe Firefly does it replace?

2 of 3 jobs. Canvas-based generative fill and expand with open models.

Text-to-image Partial replacement

Image editing (generative fill and expand) Partial replacement

Style and image reference Not covered

For style and image reference, see other Adobe Firefly alternatives. You may also need to build an interface, an API server and hosting yourself.

Facts, with sources

Every answer links to where we found it. We read the documentation; we did not install it.

Report a correction

QuestionAnswerSourceChecked

Runs on my machine? Yes, local web server and UI Invoke runs a locally hosted web server & React UI github.com · InvokeAI#readme ↗ 1d ago

Code licence Apache-2.0 github.com · LICENSE ↗ 1d ago

Model weights licence No single default model; licence depends on the model you install SD 1.5, SD 2.0, SDXL, SD 3.5 Medium, SD 3.5 Large, ... Flux.1 Dev, Flux.1 Schnell, ... Z-Image Turbo ... github.com · InvokeAI#model-support ↗ 1d ago

Commercial use Apache-2.0 code; weights vary (SD3.5 community licence, FLUX.1-dev non-commercial) SD 3.5 Large is tagged license_name: stabilityai-ai-community; FLUX.1-dev is tagged flux-1-dev-non-commercial-license on Hugging Face. huggingface.co · stable-diffusion-3.5-large ↗ 1d ago

Paid services needed? None required; some models are API-only (GPT Image, Nano Banana, Wan) Nano Banana (API Only) - GPT Image (API Only) - Wan (API Only) github.com · InvokeAI#model-support ↗ 1d ago

Hardware SD1.5 4GB VRAM/8GB RAM; SDXL 8GB/16GB; FLUX.1 10GB/32GB SD1.5 ... Nvidia 10xx+ ... 4GB ... 8GB; SDXL ... Nvidia 20xx+ ... 8GB ... 16GB; FLUX.1 ... Nvidia 20xx+ ... 10GB ... 32GB. CPU-only is described for Linux ARM64 devices and called slow. invoke.ai · system-requirements ↗ 1d ago

Operating systems Windows 10+, macOS 14+ (Apple Silicon), Linux Invoke runs on Windows 10+, macOS 14+, and Linux. AMD GPUs are supported on Linux only. invoke.ai · system-requirements ↗ 1d ago

Setup effort Easy, official Invoke Launcher The Invoke launcher is the official launcher to install, update and manage your invoke installation. invoke.ai · installation ↗ 1d ago