Skip to content

CivilisAI/Robothon-starter

 
 

Repository files navigation

🤖 Robothon 2026

Faraday Future · MuJoCo Robotics Simulation Hackathon

English · 中文

Team up with an AI coding agent, build a runnable robot simulation in MuJoCo, and submit it as a Pull Request. Every entry is scored by an AI judge panel.


📌 Key Links

📝 Register / get your contest UUID robothon.ff.com
📜 Official rules robothon.ff.com/official-rules
💬 Questions & support Discord

🚀 What is Robothon 2026?

Robothon 2026 is Faraday Future's open robotics-simulation hackathon. You partner with an AI coding tool to design and build a runnable MuJoCo robot simulation — a task, an interactive system, or a data-collection environment — and submit it as a Pull Request to this repository. Submissions are reviewed and scored entirely by an AI judge panel against one public rubric.

You don't need to be a robotics expert. Bring an idea; let the AI help you build it.


🗺️ How to Participate

  1. Register and get your contest UUID. Sign up on the official Robothon platform and copy the registration UUID issued to you.
  2. Pick an AI coding tool — Cursor, Claude Code, Kimi, Trae, or any agent you like.
  3. Fork this repository (Faraday-Future-AI/Robothon-starter), clone your fork locally (git clone https://github.com/<your-github-username>/Robothon-starter.git && cd Robothon-starter), and let your AI set up the run environment.
  4. Propose an idea, and have the AI build a runnable robot simulation in MuJoCo.
  5. Open a Pull Request, putting the same UUID in both registration.json and the PR description.
  6. Three AI judges (GPT · Claude · Gemini) score every entry independently; winners are announced and prizes awarded.
flowchart TD
    S1["1 · Register and get your contest UUID"] --> S2["2 · Pick an AI coding tool<br/>Cursor · Claude Code · Kimi · Trae …"]
    S2 --> S3["3 · Fork this repository<br/>clone locally + let AI set up"]
    S3 --> S4["4 · Propose an idea<br/>AI builds a runnable MuJoCo robot sim"]
    S4 --> S5["5 · Open a Pull Request<br/>same UUID in registration.json + PR description"]
    S5 --> S6["6 · 3 AI judges score independently<br/>GPT · Claude · Gemini"]
    S6 --> S7["🏆 Winners announced · prizes awarded"]
Loading

💡 Tip: When you hit an error, paste the message (or a screenshot) back to your AI tool and ask it to fix it — most issues resolve in a round or two.


🔑 Registration UUID (Required)

You must include the same UUID in both places:

1. In your submission folder — registration.json:

{
  "uuid": "00000000-0000-0000-0000-000000000000",
  "participant_name": "Your Name or Team Name",
  "project_name": "Your Project Name"
}

2. In your Pull Request description (the PR template will prompt you). If the template isn't shown, add this line at the top:

Registration UUID: 00000000-0000-0000-0000-000000000000

⚠️ The UUID in registration.json and your PR description must match exactly. Do not share or reuse another participant's UUID. Submissions without a valid UUID in both places may be rejected.

Use submissions/SUBMISSION_TEMPLATE/ as your starting point.


✅ Eligibility

To enter, participants must meet all of the following requirements:

  • Must be 18 or older (or the age of majority in your country/region, whichever is greater).
  • Must not be a resident of Cuba, Iran, North Korea, Syria, or Ukraine.
  • Must not be an employee, officer, director, contractor, or agent of Faraday Future and its affiliates, or an immediate family / household member of any such person.

The contest is open to eligible participants worldwide; void where prohibited or restricted by law.

See the official rules for the complete and binding terms.


🏆 Prizes & Judging

  • AI judge panel. Every submission is scored entirely by an AI panel — GPT · Claude · Gemini — with no human scoring, against one public rubric.
  • Winner determination. The winner is determined purely by the highest rubric score.
  • Prizes. Prizes will be awarded to the top entries. See the official rules for prize details.

Scoring rubric:

Criterion What we look for
Reproducibility Does the code run cleanly and is it easy to reproduce?
MuJoCo depth Use of MJCF, physics, collisions, joints, sensors, actuators
Task design Clarity, challenge, and real-world relevance
Control Teleoperation, autonomy, policy control, planning, or data collection
Dexterity Multi-finger coordination and fine manipulation (if applicable)
Engineering quality Code structure, docs, configuration, asset management
Presentation Demo-video clarity and persuasiveness
Innovation Novelty in scene, robot, task, or application

⚙️ Quick Start

git clone https://github.com/Faraday-Future-AI/Robothon-starter.git
cd $(basename Faraday-Future-AI/Robothon-starter)
python3 -m pip install -r requirements.txt

Run the example demos:

python examples/run_ff_master_demo.py
python examples/run_aegis_demo.py
python examples/run_futurist_demo.py --check-assets

run_futurist_demo.py generates a MuJoCo showcase video when the Futurist mesh files are present. Use --check-assets first to verify that every mesh referenced by assets/Futurist/futurist.urdf is included.

Open the MuJoCo viewer:

python -m mujoco.viewer

📦 What's in This Repo

Path Description
assets/Master/ FF Master humanoid MuJoCo assets (ultra / hand / fist variants)
assets/Aegis/ Aegis quadruped URDF / MuJoCo model
assets/Futurist/ FF Futurist humanoid URDF asset package
examples/ Example run scripts
model_catalog.json Reference list of recommended open-source robot models
submissions/SUBMISSION_TEMPLATE/ Submission folder template with UUID placeholder

Example scripts

Script Asset Output
examples/run_ff_master_demo.py assets/Master/scene.xml FF Master showcase video + trajectory JSON
examples/run_aegis_demo.py assets/Aegis/urdf/Aegis_mujoco.urdf Aegis patrol video + trajectory JSON
examples/run_futurist_demo.py assets/Futurist/futurist.urdf Futurist showcase video + trajectory JSON

📝 Submission Checklist

Each Pull Request should include:

  • Your project under submissions/<your-project-name>/
  • Project source code
  • MuJoCo scene files / robot models / related assets
  • Run instructions: dependencies, install steps, launch commands, controls
  • A demo video (or video link)
  • registration.json with your platform-issued UUID
  • The same UUID in your PR description
  • A short project summary: name, robot platform, task goal, technical approach, core features, highlights, current limitations, future improvements

🎥 Demo Video Requirements

The video must be produced by running your submitted code, and should show:

  • Simulation startup
  • Robot platform and task scene
  • Task execution
  • Teleoperation, autonomous control, or data-collection logic
  • Final result or task state

Recommended length: 1–3 minutes.


💡 Recommended Directions

  • Advanced teleoperation — keyboard, gamepad, VR, Web UI, motion capture
  • Long-horizon tasks — navigation, grasping, carrying, assembly, door opening, tidying, cleaning
  • Data collection — auto-generated trajectories, states, actions, images, depth, sensor streams, labels
  • Dexterous manipulation — multi-finger grasping, in-hand rotation, tool use, button presses, bottle opening
  • Real-world scenarios — K12 education, campus security, home service, warehouse logistics, industrial inspection
  • Open exploration — any creative MuJoCo robotics simulation project

Encouraged robot platforms (open-source models welcome): Unitree Go1 / Go2 / G1, Boston Dynamics Spot, Franka Emika Panda, Shadow Hand, LEAP Hand, Robotiq Gripper, or any MuJoCo / MJCF open model. See model_catalog.json, the MuJoCo Menagerie, and the MuJoCo Model Gallery.


📜 Official Rules & Legal

This page is a friendly summary. The official rules are the complete and binding terms — please read them before entering.

  • Intellectual property. You keep ownership of your submission. By entering, you grant Faraday Future a non-exclusive, royalty-free, worldwide license to use, reproduce, display, and distribute your entry for promotional and business purposes.
  • Eligibility & taxes. See the official rules for full eligibility, prize, and tax terms.
  • Sponsor. Faraday Future Intelligent Electric, Inc. (d/b/a Faraday Future), 1990 E. Grand Ave., El Segundo, CA 90245.

Nothing in this repository constitutes investment advice or a recommendation regarding any security.


Ready? Register, fork, build, and open your Pull Request. Good luck! 🚀

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Python 60.9%
  • CMake 39.1%