DimOS

DimOS is the Dimensional Framework — a modern operating system for generalist robotics (661 stars on GitHub, 135 forks, active development). Build physical applications entirely in Python that run on any humanoid, quadruped, drone, or arm.

74Trust Medium
by dimensionalOSautomationadvancedv0.0.11updated Mar 12, 2026
2.8kTotal Runs
83.0%Success Rate
661Installs
74Trust Score

Tags

#robotics#agents#mcp#hardware#humanoid#quadruped#drone#python#slam#navigation

Required Tools

bash

Inputs

NameTypeDescriptionReq
blueprinttextThe blueprint to run: e.g. "unitree-go2", "unitree-g1-sim", "unitree-go2-agentic-mcp", "drone-basic", "xarm-perception-agent".yes
commandtextCLI command: "run", "status", "stop", "log", "agent-send", "mcp list-tools", "mcp call". Defaults to "run".--
agent_messagetextNatural language message to send to the running robot agent.--
robot_iptextIP address of the real robot for hardware control.--
modetextExecution mode: "replay" (recorded data), "simulation" (MuJoCo), "live" (real hardware). Defaults to "replay".--

Outputs

NameTypeDescriptionReq
resultjsonJSON output from the CLI command.yes

Compatible Skills

SKILL.md

---
name: dimos
description: The agentive operating system for generalist robotics. Control humanoids, quadrupeds, drones, and arms with Python. No ROS required.
---

# DimOS — The Agentive Operating System for Robotics

Dimensional is the modern operating system for generalist robotics.

## Installation

### Interactive Install

```bash
curl -fsSL https://raw.githubusercontent.com/dimensionalOS/dimos/main/scripts/install.sh | bash
```

### Python Install

```bash
uv venv --python "3.12"
source .venv/bin/activate
uv pip install 'dimos[base,unitree]'
```

## Quick Start

### Run in Simulation

```bash
# Install with simulation support
uv pip install 'dimos[base,unitree,sim]'

# Run quadruped in MuJoCo simulation
dimos --simulation run unitree-go2

# Run humanoid in simulation
dimos --simulation run unitree-g1-sim
```

### Run on Real Hardware

```bash
# Control a real robot (Unitree quadruped over WebRTC)
export ROBOT_IP=<YOUR_ROBOT_IP>
dimos run unitree-go2
```

### Replay Mode (No Hardware)

```bash
# Replay a recorded quadruped session
dimos --replay run unitree-go2

# Temporal memory replay
dimos --replay --replay-dir unitree_go2_office_walk2 run unitree-go2-temporal-memory
```

## Agent CLI & MCP

```bash
# Start agentic MCP server in background
dimos run unitree-go2-agentic-mcp --daemon

# Check status
dimos status

# Follow logs
dimos log -f

# Send natural language commands
dimos agent-send "explore the room"

# List MCP tools
dimos mcp list-tools

# Call skill directly
dimos mcp call relative_move --arg forward=0.5

# Shut down
dimos stop
```

## Supported Hardware

### Quadruped
- Unitree Go2 (pro/air) 🟩 stable
- Unitree B1 🟥 experimental

### Humanoid
- Unitree G1 🟨 beta

### Arm
- Xarm 🟨 beta
- AgileX Piper 🟨 beta

### Drone
- MAVLink 🟧 alpha
- DJI Mavic 🟧 alpha

🟩 stable 🟨 beta 🟧 alpha 🟥 experimental

## Featured Blueprints

| Command | Description |
|---------|-------------|
| dimos --replay run unitree-go2 | Quadruped navigation replay |
| dimos --simulation run unitree-go2-agentic-mcp | Agentic + MCP in sim |
| dimos --simulation run unitree-g1 | Humanoid simulation |
| dimos --replay run drone-basic | Drone telemetry replay |
| dimos run demo-camera | Webcam demo (no hardware) |

## Capabilities

- **Navigation & Mapping**: SLAM, obstacle avoidance, route planning
- **Perception**: Detectors, 3D projections, VLMs, audio processing
- **Agentive Control**: Natural language robot control via LLM agents
- **Spatial Memory**: Spatio-temporal RAG, object localization
- **MCP Support**: Model Context Protocol for agent integration

## Python API

```python
from dimos.core.blueprints import autoconnect
from dimos.robot.unitree.go2.connection import go2_connection
from dimos.agents.agent import agent

blueprint = autoconnect(
    go2_connection(),
    agent(),
)

blueprint.build().loop()
```

## Resources

- GitHub: https://github.com/dimensionalOS/dimos
- Docs: AGENTS.md for detailed agent instructions
- Discord: https://discord.gg/dimos