Black Forest Labs

New release: FLUX Virtual Try-On. Sub‐4 second generations across thousands of products at low cost and low latency. We optimized for keeping a person's identity consistent and the garment’s logos, stitching, and prints all intact. Try the demo: https://lnkd.in/e3cFGjf8 Available on the BFL API: https://lnkd.in/edZYnqhS

How to use the FLUX MCP. Stephen Batifol talks through quick setup and examples: https://lnkd.in/grJNFkG3 FLUX MCP is designed to understand which model in the FLUX family to use for your request without you needing to be a prompt expert. eg it uses Klein for sub-second exploration then Pro for high-quality assets.

New release: Erase Remove objects, erase text, and clean up details in any image. FLUX is trained to erase and reconstruct as one task at the model level, this is what keeps results clean and consistent. Most removal tools leave halos, broken backgrounds, or need manual cleanup. FLUX Erase doesn't. Give it an image and a mask over what you want gone. It handles the background, shadows, and edges automatically. Available on the BFL API: https://lnkd.in/e67ZsQAC Try the demo: https://lnkd.in/eH7buSf9

The official FLUX MCP. Create images inside Claude, Cursor, Hermes, and more. Your agent automatically routes your request to the right FLUX model with expert-level prompting. From Klein for sub-second drafts to Pro for production-quality assets. Play with website visuals, ad creatives, social content, product renders, concept art, and (most importantly) memes. Link in the comments.

New release: Outpainting (expand any image to any aspect ratio). Most outpainting tools struggle with seams, prompt dependency, and rigid aspect-ratio workflows. The outputs have visible boundaries and inconsistency at expansion edges. FLUX Outpainting is a purpose-built image expansion endpoint. You provide an image, canvas size, and placement, and it generates a smooth extension carrying lighting, texture, depth, and composition through. It supports direct canvas control and up to 4MP output for production use. Available via the BFL API: https://lnkd.in/e6F_pguZ Try the demo: https://lnkd.in/em7kx2_p

From FLUX.1 to interactive visual engines. Over the last year, we’ve been focused on one question: What happens when generative models move beyond image generation into real-time visual intelligence? At AI Engineer, Stephen Batifol covers: • The evolution from FLUX.1 → FLUX.2 • Multi-reference consistency and real-time editing • Self flow: our latest self-supervised multimodal research • Improvements in text rendering, anatomy, and video consistency • The path toward scalable visual intelligence and world models, with applications in physical AI and multimodal generation We believe the future interface for creation is interactive, multimodal, and real-time. Links in the comment section.

Meet Black Forest Labs' AI Solutions team. We wrote about how our forward deployed team is working between model systems and real-world use cases. Turning model capabilities into reliable, production-grade systems is where a lot of work is happening right now. We're working closely with customers on everything from prompt design and inference pipelines to fine-tuning, evaluation, and infrastructure. If you’re building visual AI applications and hitting quality or scale challenges, feel free to reach out. Also, we're hiring.