> ## Documentation Index
> Fetch the complete documentation index at: https://agno-v2-codex-docs-audit-20260719-0149.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# GibsonAI MCP Server

> Connect a local GibsonAI MCP server to an Agno agent.

GibsonAI MCP Server - Create and manage databases with prompts

````python gibsonai.py theme={null}
"""GibsonAI MCP Server - Create and manage databases with prompts

This example shows how to connect a local GibsonAI MCP to Agno agent.
You can instantly generate, modify database schemas
and chat with your relational database using natural language.
From prompt to a serverless database (MySQL, PostgresQL, etc.), auto-generated REST APIs for your data.

Example prompts to try:
- "Create a new GibsonAI project for my e-commerce app"
- "Show me the current schema for my project"
- "Add a 'products' table with name, price, and description fields"
- "Create a 'users' table with authentication fields"
- "Deploy my schema changes to production"

How to setup and run:

1. Install [UV](https://docs.astral.sh/uv/) package manager.
2. Install the GibsonAI CLI:
    ```bash
    uvx --from gibson-cli@latest gibson auth login
    ```
3. Install the required dependencies:
    ```bash
    uv pip install agno mcp openai
    ```
4. Export your API key:
    ```bash
    export OPENAI_API_KEY="your_openai_api_key"
    ```
5. Run the GibsonAI agent by running this file.
6. Check created database and schema on GibsonAI dashboard: https://app.gibsonai.com

This logs you into the [GibsonAI CLI](https://docs.gibsonai.com/reference/cli-quickstart)
so you can access all the features directly from your agent.

"""

import asyncio
from textwrap import dedent

from agno.agent import Agent
from agno.models.openai import OpenAIChat
from agno.tools.mcp import MCPTools

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------


async def run_gibsonai_agent(message: str):
    """Run the GibsonAI agent with the given message."""
    mcp_tools = MCPTools(
        "uvx --from gibson-cli@latest gibson mcp run",
        timeout_seconds=300,  # Extended timeout for GibsonAI operations
    )

    # Connect to the MCP server
    await mcp_tools.connect()

    agent = Agent(
        name="GibsonAIAgent",
        model=OpenAIChat(id="gpt-4o"),
        tools=[mcp_tools],
        description="Agent for managing database projects and schemas",
        instructions=dedent("""\
            You are a GibsonAI database assistant. Help users manage their database projects and schemas.

            Your capabilities include:
            - Creating new GibsonAI projects
            - Managing database schemas (tables, columns, relationships)
            - Deploying schema changes to hosted databases
            - Querying database schemas and data
            - Providing insights about database structure and best practices
        """),
        markdown=True,
    )

    # Run the agent
    await agent.aprint_response(message, stream=True)

    # Close the MCP connection
    await mcp_tools.close()


# Example usage
# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------

if __name__ == "__main__":
    asyncio.run(
        run_gibsonai_agent(
            """
            Create a database for blog posts platform with users and posts tables.
            You can decide the schema of the tables without double checking with me.
            """
        )
    )
````

## Run the Example

<Steps>
  <Snippet file="create-venv-step.mdx" />

  <Step title="Install dependencies">
    ```bash theme={null}
    uv pip install -U "agno[mcp]" openai
    ```
  </Step>

  <Step title="Prepare uvx">
    Install uv, then verify `uvx` is available:

    ```bash theme={null}
    uvx --version
    ```
  </Step>

  <Step title="Export your OpenAI API key">
    <CodeGroup>
      ```bash Mac/Linux theme={null}
      export OPENAI_API_KEY="your_openai_api_key_here"
      ```

      ```bash Windows theme={null}
      $Env:OPENAI_API_KEY="your_openai_api_key_here"
      ```
    </CodeGroup>
  </Step>

  <Step title="Authenticate GibsonAI">
    Install the Gibson CLI and authenticate before starting its MCP server:

    ```bash theme={null}
    uvx --from gibson-cli@latest gibson auth login
    ```
  </Step>

  <Step title="Run the example">
    Save the code above as `gibsonai.py`, then run:

    ```bash theme={null}
    python gibsonai.py
    ```
  </Step>
</Steps>

Full source: [cookbook/91\_tools/mcp/gibsonai.py](https://github.com/agno-agi/agno/blob/v2.7.4/cookbook/91_tools/mcp/gibsonai.py)