Democratizing Embodied AI Research
We believe embodied AI should not be gatekept by expensive hardware, complex middleware, or closed ecosystems. 3we exists to remove three barriers: prohibitive costs, ROS2 complexity, and the sim-to-real gap.
Accessibility
AI researchers shouldn't need to learn ROS2. Our Python API reduces 60 lines of boilerplate to 5, letting you focus on the science.
Affordability
A research-capable robot for under $500. Complete open BOM with supplier links. No proprietary lock-in, no hidden costs.
Reproducibility
Same code runs in simulation and on hardware — guaranteed. Standardized benchmarks and public leaderboard ensure verifiable results.
Contributors
An open-source project built by researchers and engineers worldwide.
Core Maintainer
Platform Architecture
Hardware Lead
PCB & Firmware
AI/ML Lead
VLM & Gymnasium
Simulation Lead
Gazebo & Isaac
And 50+ contributors from 12 countries across academia and industry.
Open-Source Licensing
Every component of 3we is open-source under permissive licenses. You can use, modify, and commercialize freely with attribution.
| Component | License | Terms |
|---|---|---|
| Firmware & Software | Apache 2.0 | Permissive. Commercial use with attribution. |
| Hardware Designs | CERN-OHL-P v2 | Open hardware. Derivatives must include attribution. |
| Documentation | CC BY-SA 4.0 | Share-alike. Can commercialize with attribution. |
Get in Touch
For research collaborations, enterprise inquiries, or community questions.