Menlo Research
Menlo Research offers Asimov, an open-source humanoid robot platform that developers build themselves from a complete kit. The 1.2-meter-tall robot includes a browser-based simulator, full CAD files, and an agent-friendly SDK, enabling teams to program and deploy AI agents in physical environments.
- Robotics
- Developer Tools
- Artificial Intelligence
- Hardware
Funding
Founders
Founder details are not available yet.
Product
Problem
Humanoid robotics development is often locked inside proprietary, closed platforms that restrict access to hardware, software, and simulation tools. Developers and researchers who want to build on or customize a humanoid face long lead times, unavailable spare parts, and limited transparency, making experimentation slow and costly.
Solution
Menlo Research provides Asimov, an open, developer-friendly humanoid platform designed to be built, programmed, and modified by the user. The Asimov 1 kit stands 1.2 m tall, weighs 35 kg, and includes all necessary components to operate a robot in one box. It comes with a browser-based Digital Asimov simulator that mirrors the real robot's movements, allowing code testing without hardware. The full bill of materials, CAD files, simulation environment, and a pre-trained walking policy are released on GitHub, fostering a community around customization and continued development. The platform is designed to connect with AI agents through an agent-friendly SDK, enabling real-world deployment of autonomous systems.
Target Audience
Primary users are developers, robotics researchers, startups, and organizations that want hands-on access to a customizable humanoid platform for programming, research, and real-world deployment.
Features
- Self-buildable humanoid kit with 25+2 actuated joints, standing 1.2 m tall and weighing 35 kg
- Free, browser-based Digital Asimov simulator that replicates real-robot physics and movement for code testing without hardware
- Full open-source release including BOM, CAD files, simulation environment, and a pre-trained walking policy on GitHub
- Agent-friendly SDK for connecting AI agents to the robot for real-world task execution
- Assembly time of 50–100 hours, enabling hands-on learning and modification
- Designed for fleet deployments and custom builds with dedicated enterprise support