Skip to main content
Try out the MCP API Gateway inside a Meshagent Room for quick, cloud-based testing. The MCP server, tool, and chatbot will be cleaned up automatically when the Room closes. Environment variables needed:
  • API_1_NAME=<name>
    (Name of your source API, e.g. “github”)
  • API_1_SWAGGER_URL=<swagger.json-url>
    (API Swagger/OpenAPI specification URL, e.g. “https://api.github.com/swagger.json”)
  • API_1_HEADER_AUTHORIZATION=<token>
    (API authorization header value, e.g. “github_token”)

​
Steps

  1. Install Meshagent CLI: 
    pip install "meshagent[all]"
  2. Sign up & authenticate:
    Follow Meshagent CLI setup docs.
  3. Start the MCP server in a Meshagent Room: 
    meshagent service test \
  --room=test \
  --role=agent \
  --image=meshagent/mcp-api-gateway:latest \
  --env MESHAGENT_PORT=8001 \
  --env API_1_NAME=<name> \
  --env API_1_SWAGGER_URL=<swagger.json-url> \
  --env API_1_HEADER_AUTHORIZATION=<token> \
  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" \
  --name=mcp-api-gateway-service-test
    This launches the MCP API Gateway in the cloud, available only in your Meshagent Room. Room/agent shutdown cleans up automatically.
  4. Start the chatbot in the same Room: 
    meshagent chatbot join \
  --room=test \
  --agent-name=mcp-api-gateway \
  --name=mcp-api-gateway \
  --toolkit=mcp-api-gateway
    • This chatbot runs the MCP API Gateway tools. You can add more toolkits if needed.
    • When you run this, you’ll get a link to the Meshagent Room.
  5. Try it out:
    Visit the Room link in your browser. Send messages to the agent to interact with the MCP Server tools.

​
Project Level Deployment (Persistent Services)

For production, deploy MCP server tools as persistent Meshagent Project Services—your agents and chatbots automatically join every new Room in the project with zero manual steps.
  1. Create the MCP server project service: 
    meshagent service create \
  --role=agent \
  --image=meshagent/mcp-api-gateway:latest \
  --env MESHAGENT_PORT=8001 \
  --env API_1_NAME=<name> \
  --env API_1_SWAGGER_URL=<swagger.json-url> \
  --env API_1_HEADER_AUTHORIZATION=<token> \
  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" \
  --name=mcp-api-gateway-service
  2. Create the Project-level chatbot service: 
    meshagent service create \
  --image="meshagent/cli:latest" \
  --port="num=9001 path=/agent liveness=/ type=meshagent.callable participant_name=mcp-api-gateway-chatbot" \
  --env="MESHAGENT_PORT=9001" \
  --name="mcp-api-gateway-chatbot-service" \
  --command="meshagent chatbot service --agent-name=mcp-api-gateway-chatbot --toolkit=mcp-api-gateway"

​
Tools Available

These tools are provided by the integrated MCP API Gateway server:
ToolShort Description
execute_apiExecute any API endpoint with custom parameters
get_api_infoGet information about available APIs and endpoints

​
Tool Details

​
execute_api

Execute any API endpoint with custom parameters.
ParameterTypeDescription
api_namestringName of the API
methodstringHTTP method (GET, POST, etc.)
pathstringAPI endpoint path
dataobject(optional) Request body data
headersobject(optional) Additional headers
paramsobject(optional) Query parameters

​
get_api_info

Get information about available APIs and their endpoints.
ParameterTypeDescription
api_namestring(optional) Name of the API (to filter)
⌘I