Skip to content
IFAI

AI tools

BFL launches FLUX.1 Tools for fine-grained image control

Published Reach 42

BFL just released FLUX.1 Tools, adding inpainting, outpainting, depth guidance, canny edges, and variation generation to its text-to-image models. The models are available as open-access weights and API endpoints, though Depth and Canny were later deprecated. It gives creators far more precision over generated scenes.

BFL launches FLUX.1 Tools for fine-grained image control
Image: blackforestlabs.ai

FLUX.1 Tools is an official suite of control models from Black Forest Labs designed to give FLUX.1 precise inpainting, outpainting, structural guidance, and restyling capabilities.

If you have been using FLUX.1 and wanted the image quality to match the surgical precision of ControlNet, this release provides those missing levers. BFL rolled out these capabilities both as proprietary API endpoints and as guidance-distilled weights for the open FLUX.1 [dev] ecosystem.

What the tools can do

The suite launched with four distinct modules targeting specific editing workflows.

FLUX.1 Fill handles inpainting and outpainting. Given an input image, a binary mask, and a text prompt, it replaces masked sections or extends borders seamlessly. According to BFL benchmark data, Fill beats Ideogram 2.0 as well as AlimamaCreative's popular FLUX-Controlnet-Inpainting model in output quality.

FLUX.1 Depth preserves structural depth cues extracted from source pictures, making it ideal for texturing changes without breaking spatial geometry. BFL reported that it outperformed Midjourney ReTexture in evaluation tests.

FLUX.1 Canny tracks high-contrast edge maps to keep fine outlines fixed while transforming content via text prompts. Both Depth and Canny were packaged as full models and lightweight LoRA variants based on FLUX.1 [dev].

FLUX.1 Redux acts as an adapter for mixing prompts with reference imagery to generate variations or prompt-driven restyling. In BFL's API, Redux pairs with FLUX1.1 [pro] Ultra to output flexible aspect ratio images up to 4 megapixels.

Access and deprecation status

While FLUX.1 Tools began as a comprehensive suite across open weights and cloud APIs, there is an important operational caveat.

Shortly after rollout, BFL deprecated the API endpoints for FLUX.1 Depth and FLUX.1 Canny. The company stated it will no longer host or maintain those endpoints in the BFL API. However, the open-weight checkpoints and LoRA files remain available on Hugging Face under the Flux Dev License for reference and local use, with inference code on GitHub. FLUX.1 Fill and FLUX.1 Redux remain fully supported in the API.

Creators can also access the surviving tools through hosted providers like fal.ai, Replicate, Together.ai, Freepik, and krea.ai.

Why it matters

For digital artists and pipeline builders, FLUX.1 Tools turns what was previously a pure one-shot prompt generator into a real production engine. Inpainting and outpainting have long been the main bottlenecks preventing modern diffusion models from fitting smoothly into commercial asset pipelines.

BFL's decision to drop API hosting for Depth and Canny while keeping Fill and Redux suggests structural conditioning may be migrating toward local community workflows, whereas inpainting and high-resolution restyling drive the bulk of commercial API volume. If you run a local ComfyUI setup, grabbing the open weights on Hugging Face is an easy win, but API builders should stick to Fill and Redux.

Sources (1)

  1. [1]
    blackforestlabs.ai(opens in a new tab)Originalblackforestlabs.ai

Topic trackerBlack Forest Labs' FLUX model series

Every story on this topic, newest first.

  1. Black Forest Labs launches FLUX Pro finetuning API
  2. Black Forest Labs adds Ultra and Raw modes to FLUX1.1 [pro]
  3. BFL launches FLUX.1 Tools for fine-grained image controlThis story
  4. Black Forest Labs drops Flux 1.1 Pro and a dev API
  5. Black Forest Labs launches FLUX.1 with $31M backing

Best of AI, daily