Flux
FLUX is a repository by Black Forest Labs containing minimal inference code for image generation and editing using their open-weight models. It solves the problem of running state-of-the-art text-to-image and image editing models locally. It is intended for developers and researchers who want to integrate or experiment with these models.
✨ Key features
- Supports multiple open-weight models: schnell, dev, Fill, Canny, Depth, Redux, Kontext, Krea.
- Includes code for text-to-image, in/out-painting, structural conditioning, image variation, and editing.
- Provides local installation with optional TensorRT support for optimized inference.
- Offers API access to all models including Pro tier via docs.bfl.ai.
- Includes usage tracking for commercial licensing compliance.
- Reference implementations for using models with tracking enabled.
🎯 Use cases
- Generate images from text prompts using FLUX.1 [schnell] or [dev].
- Perform inpainting or outpainting with FLUX.1 Fill [dev].
- Apply structural conditioning with Canny or Depth models.
- Create image variations with FLUX.1 Redux [dev].
- Edit images in context with FLUX.1 Kontext [dev].
📦 Installation
🧰 Requirements: Python 3.10, git, and optionally enroot for TensorRT installation. For usage tracking, a BFL API key is required.
cd $HOME && git clone https://github.com/black-forest-labs/flux
cd $HOME/flux
python3.10 -m venv .venv
source .venv/bin/activate
pip install -e ".[all]"
For TensorRT support:
cd $HOME && git clone https://github.com/black-forest-labs/flux
enroot import 'docker://$oauthtoken@nvcr.io#nvidia/pytorch:25.01-py3'
enroot create -n pti2501 nvidia+pytorch+25.01-py3.sqsh
enroot start --rw -m ${PWD}/flux:/workspace/flux -r pti2501
cd flux
pip install -e ".[tensorrt]" --extra-index-url https://pypi.nvidia.com
🚀 Usage
python -m flux kontext --track_usage --prompt "replace the logo with the text 'Black Forest Labs'"
For a loop with tracking:
python -m flux kontext --track_usage --loop
⚠️ Good to know
Some models are under non-commercial licenses; commercial use requires licensing and usage tracking. TensorRT installation requires NVIDIA PyTorch container.
❓ FAQ
What models are available in this repository?
The repository includes open-weight models: FLUX.1 [schnell], [dev], Fill [dev], Canny [dev], Depth [dev], Redux [dev], Kontext [dev], and Krea [dev].
How do I install FLUX locally?
Clone the repo, create a Python 3.10 virtual environment, and run pip install -e ".[all]".
Can I use the models commercially?
Some models are under non-commercial licenses. For commercial use, you must license them via bfl.ai and enable usage tracking with a BFL API key.
How do I enable usage tracking?
Set the BFL_API_KEY environment variable and use the --track_usage flag in the CLI commands.
📊 Repository
🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.