Enquire about Robot

Simply fill up the form, our team will then reach out to you to connect to the relevant parties who can help facilitate your request.

Brand & Model
All fields required. Refer to our privacy policy for questions.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
HomeRobots

AgiBot X1

AgiBot X1 is a 1.3 m, 33 kg open-source bipedal humanoid with 34 DoF and PowerFlow actuators, designed for embodied AI research, locomotion development, and light manipulation tasks.
Software Type
Open Source
Software Package
AimRT middleware: high-performance, ROS2-compatible real-time communication framework, available on GitHub. REF-CLI: actuator configuration, diagnostics, and firmware management tool. Ubuntu 22.04 with real-time (RT) kernel on x86 host. Reinforcement learning locomotion modules available as part of the open-source repository.
Aparobot Readiness Score
ARS?
0
...
Actuators
Uses proprietary PowerFlow R series actuators (CAN bus, firmware configurable) Modular, high-precision servo motors.
Compiute
Requires external PC (x86 CPU, Ubuntu 22.04 RT). Onboard microcontrollers for actuator control.
Sensors
Equipped with joint encoders and sensors integrated within the PowerFlow actuators, the robot supports precise position and torque sensing necessary for stable locomotion and manipulation.
Max Op. time
120
mins

Recent Robot Videos

*Aparobot claims no ownership of videos posted unless otherwise stated.
Description text
Date

Robot Brief

The AgiBot X1 is a full-stack open-source bipedal humanoid robot developed by AgiBot and released in 2024 as an accessible research and development platform. Designed around a modular hardware and software architecture, the X1 is intended to lower the barrier to entry for developers, researchers, and academic institutions seeking a capable humanoid platform with full access to joint-level control and locomotion systems. Standing 1,300 mm tall and weighing 33 kg, it features 34 active degrees of freedom and is built around AgiBot's proprietary PowerFlow R series actuators, all running on a fully open-source middleware stack with ROS2 compatibility.

Use Cases

  • Bipedal Locomotion Research:
    Provides a platform for developing and validating walking controllers, balance algorithms, and reinforcement learning-based locomotion policies on a physical full-body humanoid system.
  • Light Manipulation Tasks:
    Performs object interaction and light assembly tasks using its single arm, which supports a maximum load of 0.5 kg, suitable for pick-and-place, inspection assistance, and structured manipulation research.
  • Open-Source Algorithm Development:
    Exposes joint-level control, full-body state data, and actuator configuration through the AimRT middleware and REF-CLI tools, enabling developers to implement and test custom algorithms directly on hardware.
  • Educational Demonstrations:
    Serves as a teaching and demonstration platform in academic robotics courses and research laboratories, with a compact and safe form factor designed for campus deployment.
  • Human-Robot Interaction Prototyping:
    Used as a base platform for developing and evaluating interaction behaviours, motion generation policies, and embodied AI frameworks in research settings.

Industries

  • Research and Education:
    The X1's fully open-source architecture and comprehensive SDK make it the primary deployment target for university robotics labs, AI research groups, and developers building and validating embodied intelligence algorithms.
  • Industrial Automation (Light Duty):
    Applicable to light assembly support, quality inspection demonstrations, and repetitive low-payload task automation in structured environments where a programmable humanoid platform is preferred over fixed automation.
  • Assist Healthcare:
    Applicable as a lightweight research and interaction prototype for studying human-robot collaboration in clinical or assistive care research contexts.

Specifications

Length
-
mm
Width
-
mm
Height (ResT)
-
mm
Height (Stand)
-
1300
mm
Height (Min)
-
1300
mm
Height (Max)
-
1300
mm
Weight (With Batt.)
-
33
kg
Weight (NO Batt.)
-
kg
Max Step Height
-
mm
Max Slope
+/-
-
°
Op. Temp (min)
-
°C
Op. Temp (Max)
-
°C
Ingress Rating
-
No items found.

Intro

The AgiBot X1 stands 1,300 mm tall and weighs 33 kg with its battery. Its 34 active degrees of freedom are distributed across the full body via AgiBot's proprietary PowerFlow R series actuators, which come in three variants for different joint requirements: the PowerFlow R86-3 (peak torque 200 Nm, rated at 48 V, rated torque 60 Nm, 1.28 kg), the PowerFlow R86-2 (peak torque 80 Nm, rated torque 20 Nm, 0.81 kg), and the PowerFlow R52 (peak torque 19 Nm, rated torque 6 Nm, 0.45 kg). Each actuator features hollow wiring, a multi-turn absolute encoder, a high-integrated Field-Oriented Control (FOC) drive, and a modular PF-Link interface for daisy-chaining across joint assemblies. Linear actuation in smaller joints is provided by the PowerFlow L28 linear actuator (maximum thrust 110 N, rated at 24 V), while joint coordination is managed by the Domain Controller Unit (DCU), which supports 1 kHz real-time communication forwarding at 100 Mbps and converts EtherCAT to up to three FDCAN (5 Mbps) channels. Up to 16 DCUs can be cascaded in hardware-synchronised configurations, with IMU interface support and OTA update capability. The single arm supports a maximum payload of 0.5 kg and is compatible with the AgiBot OmniPicker adaptive gripper (maximum clamping force 30 N, stroke 120 mm, 0.7-second open and close cycle).

The X1 runs on Ubuntu 22.04 with a real-time kernel on an external x86 host PC, which functions as the primary compute platform for motion planning and AI inference. Onboard microcontrollers manage actuator control loops in real time. The AimRT open-source middleware provides a high-performance ROS2-compatible communication framework, and the REF-CLI tool enables actuator firmware configuration, diagnostics, and calibration from the host system. The battery provides approximately 2 hours of runtime at a standard walking speed of 1 m/s. Full hardware documentation, URDF models, and the open-source codebase are publicly available through the AgiBot documentation portal.

Connectivity

  • USB-C data cable for communication and control
  • WiFi connectivity (2.4 GHz and 5 GHz bands assumed)
  • Actuator Bus: EtherCAT to FDCAN via DCU (1 kHz, 100 Mbps EtherCAT; 5 Mbps FDCAN per channel)
  • Peripheral Interfaces (via DCU): SPI, UART, GPIO expansion
  • AimRT middleware (ROS2 compatible); REF-CLI for actuator configuration
  • OTA Supported on all DCU-connected actuators

Capabilities

  • Open Joint-Level Control:
    Full access to all 34 joint states, torques, and velocities through the AimRT middleware and SDK, enabling low-level control for reinforcement learning policy deployment and custom locomotion research.
  • PowerFlow Actuator Ecosystem:
    Three actuator variants cover different torque and speed requirements across the body, with peak joint torques reaching 200 Nm in the highest-specification variant, enabling robust locomotion and recovery behaviours.
  • ROS2 Compatible Middleware (AimRT):
    The open-source AimRT framework provides a real-time communication backbone compatible with the broader ROS2 ecosystem, supporting sensor fusion, state estimation, and third-party module integration.
  • OTA Firmware Updates via DCU:
    The Domain Controller Unit supports over-the-air actuator firmware updates across all cascaded joints, enabling hardware upgrades without manual intervention at the joint level.
  • Modular End-Effector Support:
    The OmniPicker adaptive gripper provides force-controlled grasping with feedforward force regulation, self-locking clamping, and communication over CAN, CANFD, RS485, and serial interfaces.
  • Simulation-to-Real Transfer:
    Compatible with standard robotic simulation environments for training locomotion and manipulation policies prior to hardware deployment, supporting the embodied AI development workflow.