Direct tool calling

The call_tool method lets you call a specific MCP tool on a Connector directly, without starting a conversation or involving the model.

This is useful when you:

Already know which tool to call and what arguments to pass
Want the raw tool output for downstream processing
Are building pipelines that chain tool calls programmatically
Want to debug or verify Connector tools before using them in conversations

For scenarios where the model picks which tools to call, use Connectors in conversations instead.

note

If a Connector requires authentication, you must complete the auth flow before calling its tools.

Call a tool directly

To call a tool, pass the Connector name (or UUID), the tool name, and the arguments the tool expects.

import asyncio
from mistralai.client import Mistral

client = Mistral(api_key="your-api-key")


async def main() -> None:
    result = await client.beta.connectors.call_tool_async(
        connector_id_or_name="my_deepwiki",
        tool_name="read_wiki_structure",
        arguments={"repoName": "sqlite/sqlite"},
    )

    # result.content is a list of content blocks (TextContent, ImageContent, etc.)
    for item in result.content:
        if hasattr(item, "text"):
            print(item.text)


asyncio.run(main())

import asyncio
from mistralai.client import Mistral

client = Mistral(api_key="your-api-key")


async def main() -> None:
    result = await client.beta.connectors.call_tool_async(
        connector_id_or_name="my_deepwiki",
        tool_name="read_wiki_structure",
        arguments={"repoName": "sqlite/sqlite"},
    )

    # result.content is a list of content blocks (TextContent, ImageContent, etc.)
    for item in result.content:
        if hasattr(item, "text"):
            print(item.text)


asyncio.run(main())

The response content array contains typed content blocks (text, image, audio, or resource) returned by the MCP server.

Information

Tool names must match exactly: use list tools to check what a Connector exposes.

Chain multiple tool calls

You can call multiple tools in sequence to build a pipeline without involving the model at each step.

import asyncio
from mistralai.client import Mistral

client = Mistral(api_key="your-api-key")


async def main() -> None:
    # First call: get the repository structure
    structure = await client.beta.connectors.call_tool_async(
        connector_id_or_name="my_deepwiki",
        tool_name="read_wiki_structure",
        arguments={"repoName": "sqlite/sqlite"},
    )
    for item in structure.content:
        if hasattr(item, "text"):
            print("Structure:", item.text[:300])

    # Second call: ask a question about the repo
    answer = await client.beta.connectors.call_tool_async(
        connector_id_or_name="my_deepwiki",
        tool_name="ask_question",
        arguments={
            "repoName": "sqlite/sqlite",
            "question": "What is the purpose of the VDBE?",
        },
    )
    for item in answer.content:
        if hasattr(item, "text"):
            print("Answer:", item.text[:300])


asyncio.run(main())

import asyncio
from mistralai.client import Mistral

client = Mistral(api_key="your-api-key")


async def main() -> None:
    # First call: get the repository structure
    structure = await client.beta.connectors.call_tool_async(
        connector_id_or_name="my_deepwiki",
        tool_name="read_wiki_structure",
        arguments={"repoName": "sqlite/sqlite"},
    )
    for item in structure.content:
        if hasattr(item, "text"):
            print("Structure:", item.text[:300])

    # Second call: ask a question about the repo
    answer = await client.beta.connectors.call_tool_async(
        connector_id_or_name="my_deepwiki",
        tool_name="ask_question",
        arguments={
            "repoName": "sqlite/sqlite",
            "question": "What is the purpose of the VDBE?",
        },
    )
    for item in answer.content:
        if hasattr(item, "text"):
            print("Answer:", item.text[:300])


asyncio.run(main())