About the challenge

FIBO represents a fundamental shift in how AI generates visual content, from text prompts to structured JSON control. Instead of struggling with prompt engineering, you now have deterministic control over every aspect of generation: camera angles, FOV, lighting, color palettes, and composition - all through structured parameters that always work the same way.

The goal of this competition is simple: play, explore, and push FIBO into new directions. Show us applications that matter for real production workflows. Surprise us with real production workflows, fun and engaging experiences and new tools.

⚡You can run FIBO however you like — locally on your own setup, implement ComfyUI nodes, or through professional tools like Nuke. Build automated workflows. Design intelligent agent flows. Have a new user experience. Or push FIBO's capabilities in totally unexpected directions.

The goal is simple: play, explore, and push FIBO into new directions. Show us applications that matter for real production workflows. Surprise us with what you make.

🛡️ FIBO text-to-image foundation model was trained exclusively with more than 1 billion fully licensed images, promising great results and full indemnity in commercial use.

Get started

  • Bria FIBO models are free to download on Hugging Face: huggingface.co/briaai

  • JSON-native generation: Use the included LLM to translate natural language prompts into structured JSON for precise control

  • Learn more about FIBO's architecture and training on our blog: bria.ai/

  • GitHub repository with code examples and JSON schemas: github.com/bria-ai

  • API Documentation with parameter reference (camera angle, color palette, FOV, etc.): docs.bria.ai

  • ComfyUI workflows and ControlNet integration guides are available in the Resources section

  • Join our Discord community for support: discord.gg/briaai

Requirements

WHAT TO BUILD

Apply FIBO in an innovative way that showcases its unique capabilities. Think in terms of what falls into the following categories — we'll pick one winner for each:

Categories:

  1. Best Overall

    • The most impressive and innovative application of Bria FIBO - showcasing JSON-native control, professional parameters in a way that defines the future of enterprise visual AI. Extra focus will be given to innovations that include HDR or a 16-bit color space.

  2. Best Controllability

    • The most creative use to demonstrate FIBO's new level of output control. It can be a new multimodal ControlNet, a professional new look for parameters like camera angle, FOV, or color palette, or any way to present output controllability.

  3. Best JSON-Native or Agentic Workflow

    • The most innovative application of FIBO's JSON capabilities - using structured prompts and the LLM translator to build automated, scalable, production-ready visual generation pipelines.

  4. Best New User Experience or Professional Tool

    • The most practical production application leveraging FIBO's pro-grade controls and disentanglement for real-world creative workflows.

WHAT TO SUBMIT

  • Choose the category or categories that best align with your project and provide a brief explanation for your choice

  • Text description that explains the features and functionality of your project

  • <3-minute demonstration video of your project showing real use cases

  • Provide a URL to your public code repository with a clear indication of FIBO model use (e.g., README with setup instructions, example outputs, API integration code)

  • If applicable, include links to any deployed demos or interactive experiences

Hackathon Sponsors

Prizes

$30,000 in prizes
Best Overall
$10,000 in cash
1 winner

Mentorship with the Bria team
Featured and promoted case study

Best Controllability
$5,000 in cash
1 winner

+ Bria API credits worth $2,000

Best JSON-Native or Agentic Workflow
$5,000 in cash
1 winner

+Featured Developer Spotlight

Best New User Experience or Professional Tool
$5,000 in cash
1 winner

+Integration consultation with Bria engineers

Impact Award/Integration Award
$5,000 in cash
1 winner

Devpost Achievements

Submitting to this hackathon could earn you:

Judges

Yair Adato

Yair Adato
CEO, Bria AI

Linoy Tsaban

Linoy Tsaban
Machine Learning Engineer AI Art, Hugging Face

Rick Stringfellow

Rick Stringfellow
Head of Visual Content, EA

Grant Farhall

Grant Farhall
Chief Product Officer, Getty Images

Jonathan Fischoff

Jonathan Fischoff
Machine Learning Engineer Team Lead, FAL

Deepti Velury

Deepti Velury
CEO, Publicis Global Production

Judging Criteria

  • Usage of Bria FIBO
    Does the project use FIBO’s strengths—JSON-native generation, pro parameters, controllability, or disentangled generation—and show what sets it apart? Is the implementation technically strong?
  • Potential Impact
    How valuable could this be for professional creative workflows? Does it solve a real production problem? Could it scale to serve enterprise needs? What's the potential impact on how professionals create visual content?
  • Innovation & Creativity
    How novel is the application or approach? Does it combine FIBO's features in unexpected ways? Does it demonstrate new possibilities for structured, controllable generation? If similar tools exist, how significantly does this project improve upon them?

Questions? Email the hackathon manager

Tell your friends

Hackathon sponsors

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.