Ros2-MCP-Web-Portal
Enables natural language command control of robots via ROS2, with a web portal for real-time visualization and interaction.
README
Overview

The ROS MCP Server with Web Portal is designed to support robots in performing complex tasks and adapting effectively to various environments by providing a set of functions that transform natural language commands, entered by a user through an LLM, into ROS commands for robot control. Furthermore, by utilizing rosbridge, it is configured to operate with both ROS and ROS2 systems, and its WebSocket-based communication enables broad applicability across diverse platforms. not only that using this MCP Server I have created a web portal that makes it super easy to go through hundreds of topics and services without typing a single command and visualzie it for easy intuition, see live camera stream wth description from LLM and ability to control robots with Web-UI.
Demo
<img src="img/portal.gif" width="1500"/>
Supported Environment
- Quadruped Simulation (successful testing on Isaac-sim+IsaacLab)
- Real Quadruped (under development)
- Humanoid Simulation + Real (In future)
1. MCP-Server Installation
uv Installation
- To install
uv, you can use the following command:
curl -LsSf https://astral.sh/uv/install.sh | sh
or
pip install uv
- Create virtual environment and activate it (Optional)
uv venv
source .venv/bin/activate
MCP Server Configuration
Option A. Install Cursor
In cursor open settings and go to Tools & Integrations and then click Add Custom MCP. Then paste the code in section 2.1

Option B:
Install Claude Desktop (For Linux installation follow claude-desktop-debian).
Run claude desktop and go to developer settings and click Add MCP Server. Then paste the code given in section 2.1

MCP Config File
Paste the code in mcp.json file
{
"mcpServers": {
"ros-mcp-server": {
"command": "uv",
"args": [
"--directory",
"/ABSOLUTE/PATH/TO/Go2_Isaac_ros2/ros-mcp-server",,
"run",
"server.py"
]
}
}
}
1.1 MCP Functions
You can find the list of functions in the MCPFUNCTIONS.md.
1.2 Launch Rosbridge Server
1. Set IP and Port to connect rosbridge.
- Open
server.pyand change yourLOCAL_IP,ROSBRIDGE_IPandROSBRIDGE_PORT. (ROSBRIDGE_PORT's default value is9090)
Run rosbridge server.
ROS 1
roslaunch rosbridge_server rosbridge_websocket.launch
ROS 2
ros2 launch rosbridge_server rosbridge_websocket_launch.xml #if you dont have it install it with sudo apt command in linux

Run main.py as per README.MD
Before running, make sure num_envs in sim.yaml has same value as NUM_ENVS in server.py
Type your instructions in chat

2 Launch Web Portal
To run web portal only prerequisite is step 3 (launch robsridge server). Other steps can be ignored
2.1 Activate the evironment
Assuming that you are running it with Isaac-simulation(support for real robot will be added later), First of all activate the same environment being used for running main.py program (isaaclab_env) as per README.MD
Note: If you want to create a new seperate environment for web portal then you need to install torch and numpy as well otherwise make sure you dont update/change them in existing environment or it will cause problems.
conda activate env_isaaclab #or activate your custom env
cd ~/Go2_Isaac_ros2/ros-mcp-server
pip install -r requirements_web_mcp.txt
python web_portal.py

2.2 Changing Portal Settings
- You can choose which services to run in portal by setting parameters to True/False in config.py
- Button location might be different depending on screen size. You can adjust it in web_portal.py from line 90 onwards
- To run the LLM description dont forget to add your API key in config.py or use Ollama for local hosting. To change prompt go to llm_describer.py
- To add a new service make a new code such as lidar.py and initialize it in data_stream.py. From here all data needed for web_portal is acquired. All data is is gathered in data_stream class whose object is then called by web_portal.
- By default an image is sent to LLM every 5 seconds for description update. You can change it in data_stream.py
- If you want to use camera and controller for a different robot make sure you change the name of topics in config.py
- To share Web-portal with many users across network, Checkout Gradio-Mcp-Server
3. Acknowledgement
MCP Control is based on ros-mcp-server (Note: my implementation has deviated quite a bit now)
LLM description inspiration from vlm-describer
7. Citation
@MISC{Suleman2025,
author = "Muhammad Suleman",
title = "Ros2-MCP-Web-Portal",
year = "2025",
url = "https://github.com/sallu-786/Ros2-MCP-Web-Portal",
note = "Version 1.0.0"
}
8. Contact
sulemanmuhammad08@gmail.com
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.