reBot-DevArm README

The project's README from github.com/Seeed-Projects/reBot-DevArm. As one markdown file: README.md.

🦾 reBot-DevArm: Open Source Robotic Arm for All Developers

reBot-DevArm Banner

License: CERN-OHL-W-2.0 License: Apache-2.0 yaohui.zhu@seeed.cc ROS Support LeRobot LeRobot

🚀 100% Open Source · Embodied AI · Full Hardware + Software Stack

📦 Build your own robotic arm · 🧠 Learn robotics · 🏭 Deploy real applications

About The reBot Arm

简体中文  |  English  |  日本語 |  français |  Español

  robotics wiki

🛒 Official Purchase

🧾 reBot Arm B601-DM

🛒 SeeedStudio Bazaar 🛒 Amazon 🛒 Aliexpress
🧾 reBot Arm B601-RS

🛒 SeeedStudio Bazaar 🛒 Aliexpress

📖 Introduction

reBot-DevArm (reBot Arm B601 DM and reBot Arm B601 RS) is a robotic arm project dedicated to lowering the barrier to learning Embodied AI. We focus on "True Open Source" — not just the code, we unreservedly open source everything:

Get Your Own reBot Arm



🚀 Try the Online Simulation

Experience the reBot Arm B601-RS MuJoCo digital twin directly in your browser—no installation required. Switch between the standard arm and AGV configurations, control the joints and TCP, view the overhead and D405 wrist cameras, and run the automated put-away and stacking demos. The initial model download may take a moment; Chrome or Edge is recommended.

▶ Launch the reBot Arm Interactive Demo

🗺️ Roadmap & Status

We are committed to continuously maintaining and adapting to mainstream robot development ecosystems. Below is our current adaptation progress and planned release schedule:

reBot Arm B601 DM

Supported Ecosystem Status Description / Estimated Release Date Related Documentation
Basic Motor Usage ✅ Completed Basic motion control and API encapsulation Damiao Technology
Open-Sourcing of the New STEP 3D Structural Parts and BOM ✅ Completed STEP files for all parts in the new version, parts BOM, and reference prices for all machined components reBot Arm B601-DM BOM
Reference for Real Machine Performance Testing ✅ Completed Performance Reference of Robotic Arm under Normal and Extreme Operating Conditions Performance Testing
Assembly Video ✅ Completed Ultra-detailed assembly steps and video Getting Started with reBot Arm B601-DM
Python SDK ✅Continuously optimized, PRs welcome One-stop integration of motor read/write and control for Robstride, Damiao, Mota, Gaoqing, Hexfellow and other motors. Getting Started with Motorbridge and Web UI
ROS2 Integration ✅ Completed ROS2 integrated reBot Arm controller with support for kinematics, trajectory planning, and gravity compensation reBot Arm B601-DM ROS2 Integration Guide
Pinocchio Integration ✅ Completed Adaptation to the Pinocchio framework, enabling forward/inverse kinematics and gravity compensation for the robotic arm Getting Started with Pinocchio for reBot Arm B601-DM and Github repo
Isaac Sim Simulation 🚧 In Progress Import USD models and enable simulated teleoperation Wiki
LeRobot Integration ✅ Completed Adaptation to the Hugging Face LeRobot training framework Getting Started with LeRobot-based reBot Arm
Depth Camera Integration ✅ Completed Visual Grasping Demonstration Based on YOLO and Depth Camera Getting Started with Visual Grasping Demo
reSpeaker Voice Integration ✅ Completed Add reSpeaker Flex 4-mic array to build a voice-driven intelligent robot arm control system with spatial awareness reBot Arm B601-DM Voice Control
Gradual Updates of the Latest Algorithms ⏳ Planned Mainstream algorithms will be updated progressively Ongoing
Launch of a Series of Completely Free Courses ⏳ Planned Mainstream algorithms will be updated progressively Ongoing

Contributions from Developers

Supported Ecosystem Authors Description / Estimated Release Date Related Documentation or Repository
ROS2 (Humble), third_party integration, URDF / rebotarm_bringup @danieldoradotalaveron-rb 1. Passive diagnostics monitor (rebotarm_monitor_ros2): /diagnostics overlay for rqt_robot_monitor, serial/CAN-aware aggregator;
2. Safe park & shutdown: Capture rest pose on connect, slow return on shutdown or /rebotarm/park to prevent sudden drop;
3. Gravity compensation (smooth stop): MIT ramp-out when exiting gravity compensation to eliminate clack, jerk and instability during pos/vel handoff;
4. Gamepad teleop with IK/FK and safety measures: Gamepad control for end-effector via IK, live robot state visualization in RViz (simulation-only test);
5. D405 eye-in-hand TF: Xacro setup under end_link in rebotarm_bringup for RViz visualization & TF only (no driver/depth/intrinsics). Mount pose adjustable via launch file, bracket calibration not completed. Teleop FK/IK uses arm-only fixend_core URDF, full xacro for RSP/RViz.
rebotarm_monitor_ros2、reBotArmController_ROS2

reBot Arm B601 RS

Supported Ecosystem Status Description / Estimated Release Date Related Documentation
Basic Motor Usage ✅ Completed Basic motion control and API encapsulation Robstride
Open-Sourcing of the New STEP 3D Structural Parts and BOM ✅ Completed STEP files for all parts in the new version, parts BOM, and reference prices for all machined components reBot Arm B601-RS BOM
Getting Started ✅ Completed Quick start of B601-RS Getting Started with reBot Arm B601-RS
Assembly Video ✅ Completed Ultra-detailed assembly steps and video reBot Arm B601-RS Assembly Video
ROS2 (Humble) ✅ Completed ROS2 integrated reBot Arm controller with support for kinematics, trajectory planning, gravity compensation and MoveIt2 reBot Arm B601-DM ROS2 Integration Guide
LeRobot Integration ✅ Completed Adaptation to the Hugging Face LeRobot training framework Getting Started with LeRobot-based reBot Arm
Pinocchio Integration ✅ Completed Adaptation to the Pinocchio framework, enabling forward/inverse kinematics and gravity compensation for the robotic arm Getting Started with Pinocchio for reBot Arm B601-DM and Github repo
Depth Camera Integration ✅ Completed Visual Grasping Demonstration Based on YOLO and Depth Camera Getting Started with Visual Grasping Demo
Embodied Agent Architecture ✅ Complete Accepts natural-language commands (e.g. "pick up the red block"), automatically plans and executes grasping. Embodied Agent Architecture Design for the reBot B601 RS Robotic Arm & Source Code
Isaac Sim Simulation ✅ Completed Import USD models and enable simulated teleoperation DLI Course & Github Repo
Gradual Updates of the Latest Algorithms ⏳ Planned Mainstream algorithms will be updated progressively On going
Launch of a Series of Completely Free Courses ⏳ Planned Mainstream algorithms will be updated progressively Ongoing

⚙️ Hardware Specifications

reBot-DevArm is designed for desktop Embodied AI applications, balancing payload capacity with flexibility.

Parameter reBot Arm B601-DM reBot Arm B601-RS
Payload 1.5kg 2.5kg
Recommended Workspace 70% arm reach workspace 70% arm reach workspace
Max Reach 767 mm 754 mm
Weight Approx. 4.5 kg Approx. 6.7 kg
Repeatability < 0.2 mm < 0.2 mm
Degrees of Freedom (DOF) 6 DOF + 1 Gripper 6 DOF + 1 Gripper
Supported Platforms/Ecosystems ROS1, ROS2, LeRobot, Pinocchio, Isaac Sim, Python SDK ROS1, ROS2, LeRobot, Pinocchio, Isaac Sim, Python SDK
Supply Voltage DC 24V DC 48V

Feedback from the community

From GEM-4: Gemma Embodied 4 Physical Assistance From Linyan Fu and Apheth D Almeida From Dhruv Diddi From Ed Henderson From Sameer
From Binh_Pham From FangTianChongHui Xense YaoLin Dong From Ed Henderson From yoshikai_man
From hei-rebot-lift From Martin Kemka From Kamil Buczyński From Daniel Dorado From Asier
Seeed reBot Arm and NVIDIA Isaac Sim-to-Real VLA Course

🧹Optional Hardware

Wirst Camera Mount

32×32 UVC Intel D435i Intel D405 & Gemini 305 Gemini 2
STEP STEP STEP STEP

Compatible with Leader Arm

Star Arm 102-LD Open to compatibility integration
View Series Hub &Github repo

DIY Finger

Soft Finger Open to compatibility integration
Comming soon
[Finger Mount(ABS/

reBot-DevArm (Seeed-Projects/reBot-DevArm)

Searching Database…

这是什么?reBot-DevArm 是一个完全开源的机械臂项目(B601-DM 和 B601-RS 两个型号),旨在降低学习具身智能(Embodied AI)的门槛,所有硬件文件、BOM 和软件都公开发布。
这只是软件,还是我也得买硬件?它是一套完整的硬件+软件组合:你需要实体机械臂(从 Seeedstudio/Amazon/Aliexpress 买套件或预组装整机,或按公开的 BOM 自己 3D 打印/加工零件),外加仓库提供的 Python SDK、ROS 集成和仿真工具。
这条机械臂怎么上手?从 Seeedstudio 购买或自制五种套件之一(电机套件、结构件套件、夹爪套件、完整套件或预组装机械臂),然后按对应型号的 Getting Started 指南进行组装、基础电机控制和软件集成(Python SDK、ROS2、LeRobot 等)。你也可以在没有任何硬件的情况下,在浏览器里试用在线 MuJoCo 仿真。
要让它转起来,我实际得买什么硬件?你需要机械臂套件本身,外加一个单独的电源(默认不附带):DM/Leader 版本用 12V 10A 电源,RS 型号推荐用 Meanwell 48V 12.5A 电源(要满性能则用 48V 25A)。
路线图上的东西都已经完成了吗?大多数核心功能(基础电机控制、ROS2、Pinocchio、LeRobot、深度相机抓取、语音控制)已标记完成,但 DM 型号的 Isaac Sim 仿真等条目仍在进行中,计划中的免费课程和持续的算法更新也尚未发布。
这是面向谁的?它面向想用一条真实、完全开放的桌面机械臂(而不是封闭的商业平台)来研究或部署具身智能的开发者、机器人从业者和学习者。
它要跑在特定的操作系统或机器人框架上吗?软件栈通过 ROS1/ROS2、Isaac Sim、Pinocchio 和 Hugging Face 的 LeRobot 面向基于 Linux 的机器人环境,并带有用于电机控制的 Python SDK;在线仿真器可在任何浏览器(Chrome 或 Edge)里运行。
DM 和 RS 两个版本的机械臂有什么区别?两者外观相同,但用的是不同品牌的电机(达妙 对 Robstride);DM 型号负载 1.5kg、臂展 767mm,RS 型号负载更强、为 2.5kg,臂展 754mm,并支持执行自然语言抓取指令的具身智能体。
机械臂能听懂语音或自然语言指令吗?可以,它集成了 reSpeaker Flex 4 麦克风阵列实现语音驱动控制,B601-RS 还带有具身智能体架构,能接受“拿起红色方块”这类自然语言指令并自动规划抓取。
它支持视觉抓取吗?支持,它包含一个视觉抓取演示,把 YOLO 目标检测和深度相机(D435i、D405、Gemini)结合起来,实现基于感知的抓取和放置。
我能控制 leader/follower 遥操作配置吗?可以,这条机械臂兼容用于遥操作的 Leader Arm(Star Arm 102-LD),并通过社区的 ROS2 贡献支持基于手柄的 IK/FK 遥操作。
Size
400.1 MB
Stars
4.4K
Updated
CommitOct 19
Age
10mo
Commits · 90 Days