Back to Agile Robots
Agile ONE S

Agile ONE S

Agile ONE S

Last updated: October 27, 2026
Locomotion Type
Biped
Stage
Pilot Delivery
Height
140 cm
Degrees of Freedom
31 DOF
Hand DOF
13 DOF
Weight
40 kg
Dual-Arm Payload
6 kg
Release Year
2026
Computing
NVIDIA Jetson Thor for sensor processing and real-time inference on the robot, supported by an Intel Core i3 N305 base computing module. The software stack is open: any models, frameworks and tools can be run and swapped, and URDF plus USD assets are provided for common simulation environments. Whole-body control coordinates arms, torso and legs.
Sensors
Head-mounted RGB-D camera for colour and depth, a waist 3D LiDAR complemented by RGB-D cameras at the waist and on the back for all-around coverage. No camera or LiDAR model is published.

Key Highlights

Agile ONE S is the open reference humanoid platform in the Agile Robots portfolio: a compact 140 cm, 40 kg body with 31 degrees of freedom excluding hands and 3 kg of payload per arm, built for research teams rather than production lines. What makes it open is the software side, not just the hardware: an NVIDIA Jetson Thor handles sensor processing and real-time inference on the robot with an Intel Core i3 N305 base module alongside it, and the platform runs whatever models, frameworks and tools a lab already uses, from open-source policies to in-house pipelines, swapping them as research evolves. URDF and USD assets bring it into common simulation environments without extra integration work. The sensing rig is built for manipulation research: an RGB-D camera in the head, a 3D LiDAR at the waist with RGB-D cameras at the waist and back for all-around coverage, and a choice of hands, either the four-finger dexterous hand at 13 degrees of freedom or a two-finger gripper at 1 degree of freedom. Whole-body control coordinates arms, torso and legs so motions stay fluid and human-like, and skills and data developed here can be carried over to the full-size Agile ONE for industrial use.

Application Scenarios

Open humanoid research and manipulation experiments, training and evaluating policies with a lab's own models and frameworks, simulation-based development through URDF and USD assets, and a research-to-industry path where skills later transfer to the full-size Agile ONE.

Subscribe to Weekly Report

Get weekly updates on humanoid robot funding, new companies, and product launches.