Quickly run the MCP PostgreSQL server and a corresponding chatbot inside a Meshagent Room for testing in the cloud. When the Meshagent Room closes, both the server and chatbot are automatically removed.
  1. Install Meshagent 
    pip install "meshagent[all]"
  2. Sign up and Authenticate
    Follow the Meshagent Getting Started Guide to log in and authenticate with your credentials.
  3. Start the MCP Server in a Room 
    meshagent service test \
  --room=test \
  --role=agent \
  --image=meshagent/mcp-postgres:latest \
  --env POSTGRES_URL=postgresql://host.docker.internal:5432/mydb \
  --env MESHAGENT_PORT=8001 \
  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" \
  --name=mcp-postgres-service-test
    • This command launches a Meshagent Room named test running a test MCP server agent.
    • Environment Variables:
      • POSTGRES_URL: The URL to your PostgreSQL instance (e.g. postgresql://host.docker.internal:5432/mydb)
      • MESHAGENT_PORT: Listening port for the MCP server (default: 8001)
    • Note: Rooms will close if inactive, cleaning up all agents and tools.
  4. Start a Chatbot with Postgres Toolkit in the Room 
    meshagent chatbot join \
  --room=test \
  --agent-name=mcp-postgres \
  --name=mcp-postgres \
  --toolkit=mcp-postgres
    • Launches a chatbot in the same room that can use the Postgres toolkit.
    • Multiple toolkits can be used by a single agent.
    • A joining link to the room is given in the command output.
  5. Interact with the tools
    • Open the room link in your browser and send messages to the mcp-postgres agent to use the MCP server tools directly.

​
Project Level Deployment

For persistent, production setups, create project-level services so that your MCP server and chatbot are available in every Meshagent project room—no manual room joining needed, no separate local hosting.
  1. Deploy MCP Server as a Persistent Project-Level Service 
    meshagent service create \
  --role=agent \
  --image=meshagent/mcp-postgres:latest \
  --env POSTGRES_URL=postgresql://host.docker.internal:5432/mydb \
  --env MESHAGENT_PORT=8001 \
  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" \
  --name=mcp-postgres-service
  2. Deploy Chatbot as a Persistent Project-Level Service 
    meshagent service create \
  --image="meshagent/cli:latest" \
  --port="num=9001 path=/agent liveness=/ type=meshagent.callable participant_name=mcp-postgres-chatbot" \
  --env="MESHAGENT_PORT=9001" \
  --name="mcp-postgres-chatbot-service" \
  --command="meshagent chatbot service --agent-name=mcp-postgres-chatbot --toolkit=mcp-postgres"
After setting these up, each Meshagent project room will automatically have access to the MCP server tool and the chatbot when created.

​
Tools Available

​
PostgreSQL readonly (Archived) MCP Server

Connect with read-only access to PostgreSQL databases. This server enables LLMs to inspect database schemas and execute read-only queries.

​
Available Tools

Tools provided by this ServerShort Description
queryRun a read-only SQL query

​
Tool Details

Tool: query
Run a read-only SQL query
Parameters:
NameTypeDescription
sqlstringoptional

​
Use this MCP Server (standalone Docker)

{
  "mcpServers": {
    "postgres": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "POSTGRES_URL",
        "mcp/postgres",
        "$POSTGRES_URL"
      ],
      "env": {
        "POSTGRES_URL": "postgresql://host.docker.internal:5432/mydb"
      }
    }
  }
}