explaingit

black-forest-labs/flux

★ 25,552PythonAudience · developerComplexity · 3/5QuietLicenseSetup · hard

TLDR

FLUX.1 is an AI image generation and editing toolkit that turns text into pictures, edits existing images, and runs locally on your computer or via API.

Mindmap

mindmap
  root((FLUX.1))
    What it does
      Text to image
      Image editing
      Image variation
      Structural conditioning
    Model variants
      Schnell fast
      Dev high quality
      Kontext editing
    How to use
      Local installation
      HuggingFace models
      Optional TensorRT
    Use cases
      Creative workflows
      Product integration
      Research projects

Things people build with this

USE CASE 1

Generate high-quality images from text prompts for creative projects, marketing, or product design.

USE CASE 2

Edit or extend existing images by filling in missing areas or adding new content to the sides.

USE CASE 3

Create variations of an image or guide generation using depth maps and edge outlines for consistent style.

USE CASE 4

Integrate AI image generation into applications or workflows without relying on external APIs.

Tech stack

PythonHuggingFacePyTorchTensorRTNVIDIA CUDA

Getting it running

Difficulty · hard Time to first run · 1h+

Requires NVIDIA GPU with CUDA, large model downloads (10GB+), and PyTorch/TensorRT compilation.

Schnell models are Apache-licensed and free for commercial use; dev models are non-commercial unless you purchase a commercial license from Black Forest Labs.

In plain English

FLUX is the official code repository for running FLUX.1, a suite of AI image generation and editing models created by Black Forest Labs. It solves the problem of generating high-quality images from text descriptions or editing existing images, all running on your own computer or via an API. The library offers several model variants for different tasks: text-to-image (turn a written prompt into a picture), in-painting and out-painting (fill in or extend parts of an image), structural conditioning (guide image style using depth maps or edge outlines from another image), image variation (create new versions of an existing image), and image editing using a model called Kontext. Models come in two main flavors, "schnell" (fast, Apache-licensed, free for commercial use) and "dev" (higher quality, non-commercial license unless separately licensed through Black Forest Labs' commercial tier). You would use this if you are a researcher, developer, or creative professional who wants to run state-of-the-art AI image generation locally or integrate it into a product. The tech stack is Python, and the models are downloaded from HuggingFace. Optional TensorRT support from NVIDIA is available for faster performance on compatible graphics cards.

Copy-paste prompts

Prompt 1
How do I install FLUX.1 and generate an image from a text prompt using the schnell model?
Prompt 2
Show me how to use in-painting with FLUX to fill in a masked area of an existing image.
Prompt 3
How can I use structural conditioning with a depth map to guide the style of generated images?
Prompt 4
What's the difference between schnell and dev models, and when should I use each one?
Prompt 5
How do I set up TensorRT acceleration for faster image generation on an NVIDIA GPU?
Open on GitHub → Explain another repo

Generated 2026-05-18 · Model: sonnet-4-6 · Verify against the repo before relying on details.