InvokeAI

6.12.0 verified Fri May 01 auth: no python

InvokeAI is a full-featured AI-assisted image generation environment designed for creatives and enthusiasts. It supports multiple diffusion models including Stable Diffusion, FLUX, Z-Image Turbo, and others, with a rich web UI, node-based workflow editor, and comprehensive model management. Current version is 6.12.0; release cadence is roughly monthly with release candidates.

pip install invokeai

Common errors

error ModuleNotFoundError: No module named 'invokeai' ↓

cause InvokeAI not installed or installed in a different environment.

fix

Run pip install invokeai and ensure the virtual environment is activated.

error ImportError: cannot import name 'InvokeAIApp' from 'invokeai' ↓

cause Incorrect import path; InvokeAIApp is in the app subpackage.

fix

Use from invokeai.app import InvokeAIApp.

error RuntimeError: model not loaded ↓

cause No model is installed or selected in the model manager.

fix

Install a model via invokeai-model-install --model runwayml/stable-diffusion-v1-5 or via the web UI.

Warnings

breaking Python 3.13 is not supported; requires Python >=3.11,<3.13. ↓

fix Use Python 3.11 or 3.12.

breaking The model directory structure changed in v6.9.0 to a flat layout. Existing setups will be migrated on first run. ↓

fix Let the migration run automatically; do not manually rearrange model folders.

gotcha InvokeAI uses its own model registry; simply placing .safetensors in a folder does not work. You must install models via the UI or CLI. ↓

fix Use `invokeai-model-install` CLI or the Model Manager in the web UI.

Imports

InvokeAIApp
wrong
```
from invokeai import InvokeAIApp
```
correct
```
from invokeai.app import InvokeAIApp
```
InvokeAIApp is in the app subpackage; direct import fails.

ModelManager

wrong

from invokeai.model_manager import ModelManager

correct

from invokeai.backend.model_manager import ModelManager

ModelManager moved to backend.model_manager in v6.x.

Quickstart

Basic usage of InvokeAI programmatically to generate an image. Requires a loaded model.

import os
from invokeai.app import InvokeAIApp

app = InvokeAIApp()
# Run a simple txt2img generation
result = app.generate(
    prompt="A serene mountain landscape",
    width=512,
    height=512,
    steps=20,
    cfg_scale=7.5
)
print("Images saved:", result.images)