Using Connectors in conversations
After a Connector is registered and authenticated (if required), you can attach it to any conversation.
You can mix Connectors with built-in tools like web_search, filter which tools the model can use, and configure Connectors on Agents so they're available in every conversation.
Tools are MCP server executable functions the model can call. When you attach a Connector, the model discovers these tools automatically and calls the right ones based on the query.
Attach Connectors to a conversation
Basic usage
Pass any Connector as a tool to a conversation using type: "connector" and the Connector's name or UUID as connector_id.
All tools the Connector exposes are available to the model. To restrict which tools the model can call, see Filter tools.
import asyncio
from mistralai.client import Mistral
client = Mistral(api_key="your-api-key")
async def main() -> None:
response = await client.beta.conversations.start_async(
model="mistral-small-latest",
inputs=[
{
"role": "user",
"content": "Using deepwiki, tell me about the structure of the sqlite/sqlite repository.",
}
],
tools=[
{
"type": "connector",
"connector_id": "my_deepwiki", # name or UUID
},
],
)
for output in response.outputs:
if output.type == "message.output":
print(output.content)
asyncio.run(main())Filter tools
To control which tools from a Connector the model can use, add a tool_configuration object. Set include to allowlist specific tools, or exclude to block them. You can use one or the other, but not both at the same time.
import asyncio
from mistralai.client import Mistral
client = Mistral(api_key="your-api-key")
async def main() -> None:
# Exclude specific tools
response = await client.beta.conversations.start_async(
model="mistral-small-latest",
inputs=[{"role": "user", "content": "What tools do you have access to?"}],
tools=[
{
"type": "connector",
"connector_id": "my_deepwiki",
"tool_configuration": {
"exclude": ["read_wiki_structure"],
},
},
],
)
# Include only specific tools
response = await client.beta.conversations.start_async(
model="mistral-small-latest",
inputs=[{"role": "user", "content": "What tools do you have access to?"}],
tools=[
{
"type": "connector",
"connector_id": "my_deepwiki",
"tool_configuration": {
"include": ["ask_question"],
},
},
],
)
asyncio.run(main())You can require user approval before a tool runs by adding requires_confirmation to the tool_configuration. This is useful for sensitive actions like sending emails or modifying data. See Human-in-the-loop for the full confirmation flow.
Use built-in tools alongside Connectors
Mistral provides built-in tools such as web_search, code_interpreter, image_generation, and document_library. You can pass them alongside Connectors in the same tools array. The model decides which tool to call based on the query.
import asyncio
from mistralai.client import Mistral
client = Mistral(api_key="your-api-key")
async def main() -> None:
response = await client.beta.conversations.start_async(
model="mistral-small-latest",
inputs=[
{
"role": "user",
"content": "Search the web for the latest SQLite release, then use deepwiki to find where the version is defined in the sqlite/sqlite repo.",
}
],
tools=[
{"type": "web_search"},
{"type": "connector", "connector_id": "my_deepwiki"},
],
)
for output in response.outputs:
if output.type == "message.output":
print(output.content)
asyncio.run(main())Add Connectors to an Agent
You can also attach Connectors directly to an Agent at creation time. Every conversation started with that Agent then has access to the Connector tools automatically, without passing a tools array each time.
This is useful when an Agent always needs the same external tools (for example, a support Agent that always queries a CRM).
import asyncio
from mistralai.client import Mistral
client = Mistral(api_key="your-api-key")
async def main() -> None:
# Create the agent with a connector
agent = await client.beta.agents.create_async(
name="deepwiki_agent",
description="Agent with DeepWiki access for code repository exploration",
model="mistral-small-latest",
instructions="You are a helpful assistant that can explore code repositories using DeepWiki. Be concise.",
tools=[
{
"type": "connector",
"connector_id": "my_deepwiki",
},
],
)
# Start a conversation using the agent — no tools needed here
response = await client.beta.conversations.start_async(
agent_id=agent.id,
inputs=[
{"role": "user", "content": "What is the main purpose of the sqlite repository?"}
],
)
for output in response.outputs:
if output.type == "message.output":
print(output.content)
asyncio.run(main())Use agent_id instead of model when starting a conversation. You cannot pass both.