> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tavily.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Mistral Studio

> Give Mistral models and agents live web search, extraction, crawling, and research through the Tavily connector in Mistral Studio.

[Mistral Studio](https://mistral.ai/products/studio/) is Mistral's platform for building, deploying, and governing agentic AI systems. Tavily is available as a managed [connector](https://docs.mistral.ai/studio/connectors/overview) in the Mistral Studio marketplace. Add it once and your Mistral agents and conversations can search the web for fresh information as they work, with no custom search integration to build or maintain.

## Add the Tavily connector

Create a [Mistral Studio](https://console.mistral.ai) account and a [Tavily account](https://app.tavily.com).

**1. Open the Tavily connector**

Open the [Tavily connector](https://console.mistral.ai/connectors/marketplace?template=tavily) directly, or in Mistral Studio go to **Context → Connectors → Marketplace** and select **Tavily**. The server is preconfigured as `https://mcp.tavily.com/mcp` and uses OAuth.

**2. Connect with Tavily**

Click **Connect**, sign in to Tavily, and return to Mistral Studio. The connector now shows as connected.

**3. Try it**

Attach the Tavily connector to an agent or conversation, then ask a question that needs current information, for example:

> Use Tavily to find three AI product announcements from the past seven days. Cite your sources.

Tavily tool calls and source links in the response confirm it worked.

You can also create a custom connector from the **Connectors** tab in Mistral Studio, for example to point at your own hosted or remote MCP server, or to authenticate with an API key or credentials shared across your workspace. See Mistral's [connector docs](https://docs.mistral.ai/studio/connectors/management).

## Use the connector from code

Reference the connector by its name, `tavily`.

### In an agent

Add the connector to a [Mistral agent](https://docs.mistral.ai/studio/agents/learn/agent/connectors):

```python theme={null}
from mistralai import agents

tavily = agents.connector("tavily")

agent = agents.Agent(
    model="mistral-large-4",
    name="web-research-agent",
    description="Answers questions from the web using Tavily.",
    instructions="Answer questions by searching the web. Cite the URLs you used.",
    connectors=[tavily],
)
```

### In a conversation

Pass the connector in the `tools` array of a [conversation](https://docs.mistral.ai/studio/connectors/conversations):

```python theme={null}
import asyncio
from mistralai.client import Mistral

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


async def main() -> None:
    response = await client.beta.conversations.start_async(
        model="mistral-small-latest",
        inputs=[
            {
                "role": "user",
                "content": "Use Tavily to find the latest stable release of Python. Cite your source.",
            }
        ],
        tools=[{"type": "connector", "connector_id": "tavily"}],
    )
    for output in response.outputs:
        if output.type == "message.output":
            print(output.content)


asyncio.run(main())
```

To limit which Tavily tools the model can call, add a `tool_configuration` with `include` or `exclude`. See [Filter tools](https://docs.mistral.ai/studio/connectors/conversations#filtering-tools).

## Troubleshooting

* **Mistral Studio says a connector for this provider already exists.** Someone in your workspace has already added Tavily. Open **View connectors** to add a connection or manage permissions on the existing one, or create a separate connector.
* **Sign-in doesn't complete.** Click **Connect** again and finish Tavily's sign-in in the browser tab Mistral Studio opens.
* **The agent doesn't use Tavily.** Confirm the Tavily connector shows as connected and is attached to the agent or conversation, then name it in your prompt, for example "Use Tavily to…".
* **OAuth uses the wrong account.** Disconnect Tavily in Mistral Studio, connect again, and choose the account you want. See [Choosing an account for OAuth](https://docs.tavily.com/documentation/mcp.md#oauth-authentication).

## Resources

* [Tavily connector in the Mistral Studio marketplace](https://console.mistral.ai/connectors/marketplace?template=tavily)
* [Mistral connectors overview](https://docs.mistral.ai/studio/connectors/overview)
* [Use connectors in conversations](https://docs.mistral.ai/studio/connectors/conversations)
* [Connectors in agents](https://docs.mistral.ai/studio/agents/learn/agent/connectors)
* [Tavily MCP server](https://docs.tavily.com/documentation/mcp.md)
