SIGN UPLOG IN
X (Twitter)LinkedInGitHubDiscordYouTube

How to integrate Mistral ai MCP with Vercel AI SDK v6

Connect Vercel AI SDK v6 to Mistral ai MCP. Summarize this research paper in simple terms, generate python code for sorting a list, and more using natural language, with authentication handled for you.

Get started for freeGet a demo
Mistral ai logoMistral ai
Api Key

Mistral ai is a research lab offering cutting-edge open-source language models and developer APIs. It empowers teams to add state-of-the-art natural language capabilities to any app or workflow.

105 Tools

Introduction

This guide walks you through connecting Mistral ai to Vercel AI SDK v6 using the Composio tool router. By the end, you'll have a working Mistral ai agent that can summarize this research paper in simple terms, generate python code for sorting a list, explain the difference between ai and ml through natural language commands.

This guide will help you understand how to give your Vercel AI SDK agent real control over a Mistral ai account through Composio's Mistral ai MCP server.

Before we dive in, let's take a quick look at the key ideas and tools involved.

Also integrate Mistral ai with

Claude CodeCursorChatGPTClaude CoworkAntigravityOpenAI Agents SDKClaude Agent SDKCodexGrok BuildVS CodeOpenCodeOpenClawHermesAtomic AgentGoogle ADKLangChainMastra AILlamaIndexCrewAIPydantic AIAutoGen

TL;DR

Here's what you'll learn:
  • How to set up and configure a Vercel AI SDK agent with Mistral ai integration
  • Using Composio's Tool Router to dynamically load and access Mistral ai tools
  • Creating an MCP client connection using HTTP transport
  • Building an interactive CLI chat interface with conversation history management
  • Handling tool calls and results within the Vercel AI SDK framework

What is Vercel AI SDK?

The Vercel AI SDK is a TypeScript library for building AI-powered applications. It provides tools for creating agents that can use external services and maintain conversation state.

Key features include:

  • streamText: Core function for streaming responses with real-time tool support
  • MCP Client: Built-in support for Model Context Protocol via @ai-sdk/mcp
  • Step Counting: Control multi-step tool execution with stopWhen: stepCountIs()
  • OpenAI Provider: Native integration with OpenAI models

What is the Mistral ai MCP server, and what's possible with it?

The Mistral ai MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Mistral ai account. It provides structured and secure access to your Mistral AI models, so your agent can perform actions like generating text, summarizing content, answering questions, extracting structured information, and handling advanced language tasks on your behalf.

  • Text generation and completion: Have your agent produce coherent, context-aware text responses, complete prompts, or generate creative content leveraging Mistral's advanced models.
  • Summarization and paraphrasing: Ask your agent to summarize lengthy documents or rephrase input text for improved clarity or brevity.
  • Question answering and information extraction: Let your agent answer questions, extract key facts, or pull structured data from unstructured content automatically.
  • Content classification and sentiment analysis: Enable your agent to categorize text, detect topics, or analyze sentiment to inform downstream workflows.
  • Conversational AI and dialogue management: Build rich, multi-turn conversations or chatbots that handle context, intent, and user queries seamlessly using Mistral's models.

What is the Composio tool router, and how does it fit here?

What is Composio SDK?

Composio's Composio SDK helps agents find the right tools for a task at runtime. You can plug in multiple toolkits (like Gmail, HubSpot, and GitHub), and the agent will identify the relevant app and action to complete multi-step workflows. This can reduce token usage and improve the reliability of tool calls. Read more here: Getting started with Composio SDK

The tool router generates a secure MCP URL that your agents can access to perform actions.

How the Composio SDK works

The Composio SDK follows a three-phase workflow:

  1. Discovery: Searches for tools matching your task and returns relevant toolkits with their details.
  2. Authentication: Checks for active connections. If missing, creates an auth config and returns a connection URL via Auth Link.
  3. Execution: Executes the action using the authenticated connection.

Step-by-step Guide

Step by step09 STEPS
1

Prerequisites

Before you begin, make sure you have:
  • Node.js and npm installed
  • A Composio account with API key
  • An OpenAI API key
2

Getting API Keys for OpenAI and Composio

OpenAI API Key
  • Go to the OpenAI dashboard and create an API key. You'll need credits to use the models, or you can connect to another model provider.
  • Keep the API key safe.
Composio API Key
  • Log in to the Composio dashboard.
  • Navigate to your API settings and generate a new API key.
  • Store this key securely as you'll need it for authentication.
3

Install required dependencies

bash
npm install @ai-sdk/openai @ai-sdk/mcp @composio/core ai dotenv

First, install the necessary packages for your project.

What you're installing:

  • @ai-sdk/openai: Vercel AI SDK's OpenAI provider
  • @ai-sdk/mcp: MCP client for Vercel AI SDK
  • @composio/core: Composio SDK for tool integration
  • ai: Core Vercel AI SDK
  • dotenv: Environment variable management
4

Set up environment variables

bash
OPENAI_API_KEY=your_openai_api_key_here
COMPOSIO_API_KEY=your_composio_api_key_here
COMPOSIO_USER_ID=your_user_id_here

Create a .env file in your project root.

What's needed:

  • OPENAI_API_KEY: Your OpenAI API key for GPT model access
  • COMPOSIO_API_KEY: Your Composio API key for tool access
  • COMPOSIO_USER_ID: A unique identifier for the user session
5

Import required modules and validate environment

typescript
import "dotenv/config";
import { openai } from "@ai-sdk/openai";
import { Composio } from "@composio/core";
import * as readline from "readline";
import { streamText, type ModelMessage, stepCountIs } from "ai";
import { createMCPClient } from "@ai-sdk/mcp";

const composioAPIKey = process.env.COMPOSIO_API_KEY;
const composioUserID = process.env.COMPOSIO_USER_ID;

if (!process.env.OPENAI_API_KEY) throw new Error("OPENAI_API_KEY is not set");
if (!composioAPIKey) throw new Error("COMPOSIO_API_KEY is not set");
if (!composioUserID) throw new Error("COMPOSIO_USER_ID is not set");

const composio = new Composio({
  apiKey: composioAPIKey,
});
What's happening:
  • We're importing all necessary libraries including Vercel AI SDK's OpenAI provider and Composio
  • The dotenv/config import automatically loads environment variables
  • The MCP client import enables connection to Composio's tool server
6

Create Tool Router session and initialize MCP client

typescript
async function main() {
  // Create a tool router session for the user
  const session = await composio.create(composioUserID!, {
    toolkits: ["mistral_ai"],
  });

  const mcpUrl = session.mcp.url;
What's happening:
  • We're creating a Tool Router session that gives your agent access to Mistral ai tools
  • The create method takes the user ID and specifies which toolkits should be available
  • The returned mcp object contains the URL and authentication headers needed to connect to the MCP server
  • This session provides access to all Mistral ai-related tools through the MCP protocol
7

Connect to MCP server and retrieve tools

typescript
const mcpClient = await createMCPClient({
  transport: {
    type: "http",
    url: mcpUrl,
    headers: session.mcp.headers, // Authentication headers for the Composio MCP server
  },
});

const tools = await mcpClient.tools();
What's happening:
  • We're creating an MCP client that connects to our Composio Tool Router session via HTTP
  • The mcp.url provides the endpoint, and mcp.headers contains authentication credentials
  • The type: "http" is important - Composio requires HTTP transport
  • tools() retrieves all available Mistral ai tools that the agent can use
8

Initialize conversation and CLI interface

typescript
let messages: ModelMessage[] = [];

console.log("Chat started! Type 'exit' or 'quit' to end the conversation.\n");
console.log(
  "Ask any questions related to mistral_ai, like summarize my last 5 emails, send an email, etc... :)))\n",
);

const rl = readline.createInterface({
  input: process.stdin,
  output: process.stdout,
  prompt: "> ",
});

rl.prompt();
What's happening:
  • We initialize an empty messages array to maintain conversation history
  • A readline interface is created to accept user input from the command line
  • Instructions are displayed to guide the user on how to interact with the agent
9

Handle user input and stream responses with real-time tool feedback

typescript
rl.on("line", async (userInput: string) => {
  const trimmedInput = userInput.trim();

  if (["exit", "quit", "bye"].includes(trimmedInput.toLowerCase())) {
    console.log("\nGoodbye!");
    rl.close();
    process.exit(0);
  }

  if (!trimmedInput) {
    rl.prompt();
    return;
  }

  messages.push({ role: "user", content: trimmedInput });
  console.log("\nAgent is thinking...\n");

  try {
    const stream = streamText({
      model: openai("gpt-5"),
      messages,
      tools,
      toolChoice: "auto",
      stopWhen: stepCountIs(10),
      onStepFinish: (step) => {
        for (const toolCall of step.toolCalls) {
          console.log(`[Using tool: ${toolCall.toolName}]`);
          }
          if (step.toolCalls.length > 0) {
            console.log(""); // Add space after tool calls
          }
        },
      });

      for await (const chunk of stream.textStream) {
        process.stdout.write(chunk);
      }

      console.log("\n\n---\n");

      // Get final result for message history
      const response = await stream.response;
      if (response?.messages?.length) {
        messages.push(...response.messages);
      }
    } catch (error) {
      console.error("\nAn error occurred while talking to the agent:");
      console.error(error);
      console.log(
        "\nYou can try again or restart the app if it keeps happening.\n",
      );
    } finally {
      rl.prompt();
    }
  });

  rl.on("close", async () => {
    await mcpClient.close();
    console.log("\n👋 Session ended.");
    process.exit(0);
  });
}

main().catch((err) => {
  console.error("Fatal error:", err);
  process.exit(1);
});
What's happening:
  • We use streamText instead of generateText to stream responses in real-time
  • toolChoice: "auto" allows the model to decide when to use Mistral ai tools
  • stopWhen: stepCountIs(10) allows up to 10 steps for complex multi-tool operations
  • onStepFinish callback displays which tools are being used in real-time
  • We iterate through the text stream to create a typewriter effect as the agent responds
  • The complete response is added to conversation history to maintain context
  • Errors are caught and displayed with helpful retry suggestions

Complete Code

Here's the complete code to get you started with Mistral ai and Vercel AI SDK:

typescript
import "dotenv/config";
import { openai } from "@ai-sdk/openai";
import { Composio } from "@composio/core";
import * as readline from "readline";
import { streamText, type ModelMessage, stepCountIs } from "ai";
import { createMCPClient } from "@ai-sdk/mcp";

const composioAPIKey = process.env.COMPOSIO_API_KEY;
const composioUserID = process.env.COMPOSIO_USER_ID;

if (!process.env.OPENAI_API_KEY) throw new Error("OPENAI_API_KEY is not set");
if (!composioAPIKey) throw new Error("COMPOSIO_API_KEY is not set");
if (!composioUserID) throw new Error("COMPOSIO_USER_ID is not set");

const composio = new Composio({
  apiKey: composioAPIKey,
});

async function main() {
  // Create a tool router session for the user
  const session = await composio.create(composioUserID!, {
    toolkits: ["mistral_ai"],
  });

  const mcpUrl = session.mcp.url;

  const mcpClient = await createMCPClient({
    transport: {
      type: "http",
      url: mcpUrl,
      headers: session.mcp.headers, // Authentication headers for the Composio MCP server
    },
  });

  const tools = await mcpClient.tools();

  let messages: ModelMessage[] = [];

  console.log("Chat started! Type 'exit' or 'quit' to end the conversation.\n");
  console.log(
    "Ask any questions related to mistral_ai, like summarize my last 5 emails, send an email, etc... :)))\n",
  );

  const rl = readline.createInterface({
    input: process.stdin,
    output: process.stdout,
    prompt: "> ",
  });

  rl.prompt();

  rl.on("line", async (userInput: string) => {
    const trimmedInput = userInput.trim();

    if (["exit", "quit", "bye"].includes(trimmedInput.toLowerCase())) {
      console.log("\nGoodbye!");
      rl.close();
      process.exit(0);
    }

    if (!trimmedInput) {
      rl.prompt();
      return;
    }

    messages.push({ role: "user", content: trimmedInput });
    console.log("\nAgent is thinking...\n");

    try {
      const stream = streamText({
        model: openai("gpt-5"),
        messages,
        tools,
        toolChoice: "auto",
        stopWhen: stepCountIs(10),
        onStepFinish: (step) => {
          for (const toolCall of step.toolCalls) {
            console.log(`[Using tool: ${toolCall.toolName}]`);
          }
          if (step.toolCalls.length > 0) {
            console.log(""); // Add space after tool calls
          }
        },
      });

      for await (const chunk of stream.textStream) {
        process.stdout.write(chunk);
      }

      console.log("\n\n---\n");

      // Get final result for message history
      const response = await stream.response;
      if (response?.messages?.length) {
        messages.push(...response.messages);
      }
    } catch (error) {
      console.error("\nAn error occurred while talking to the agent:");
      console.error(error);
      console.log(
        "\nYou can try again or restart the app if it keeps happening.\n",
      );
    } finally {
      rl.prompt();
    }
  });

  rl.on("close", async () => {
    await mcpClient.close();
    console.log("\n👋 Session ended.");
    process.exit(0);
  });
}

main().catch((err) => {
  console.error("Fatal error:", err);
  process.exit(1);
});

Conclusion

You've successfully built a Mistral ai agent using the Vercel AI SDK with streaming capabilities! This implementation provides a powerful foundation for building AI applications with natural language interfaces and real-time feedback.

Key features of this implementation:

  • Real-time streaming responses for a better user experience with typewriter effect
  • Live tool execution feedback showing which tools are being used as the agent works
  • Dynamic tool loading through Composio's Tool Router with secure authentication
  • Multi-step tool execution with configurable step limits (up to 10 steps)
  • Comprehensive error handling for robust agent execution
  • Conversation history maintenance for context-aware responses

You can extend this further by adding custom error handling, implementing specific business logic, or integrating additional Composio toolkits to create multi-app workflows.
TOOLS

Supported Tools

Every Mistral ai action and event your agent gets out of the box.

Append to conversation

Tool to append new entries to an existing conversation in Mistral AI.

Archive Fine-Tuned Model

Archive a fine-tuned model so it is unavailable until restored.

Count Chat Completion Field Options

Count values of a chat completion field after optional filtering.

Create Agent

Tool to create a new AI agent with custom configuration (Beta).

Create Agents Completion

Tool to generate completions using a Mistral AI agent with specific instructions and tools.

Create Audio Transcription

Transcribe audio files to text using Mistral AI's Voxtral models.

Create Batch Job

Create and queue a Mistral batch inference job.

Create Chat Completion

Generate conversational responses from Mistral AI models.

Create Chat Moderation

Tool to classify chat content for moderation purposes across 9 categories.

Create Embeddings

Tool to generate vector embeddings for input text using Mistral AI embedding models.

Create FIM Completion

Generate code completions using fill-in-the-middle functionality.

Create library

Tool to create a new document library.

Create library share

Create or update sharing permissions for a library.

Create Moderation

Tool to classify text content for moderation purposes across 9 categories.

Create OCR

Extract text and structured data from images and documents using Mistral AI's OCR capabilities.

Create or Update Agent Alias

Tool to create or update an agent version alias.

Create Speech

Generate JSON-encoded speech audio from text without streaming.

Delete agent

Permanently deletes an agent by its ID (Beta feature).

Delete Conversation

Tool to delete a conversation by its ID (Beta).

Delete File

Delete a file by its ID from Mistral AI.

Delete library

Permanently deletes a library and all of its documents from Mistral AI.

Delete library document

Permanently deletes a document from a Mistral AI library.

Delete library share

Remove sharing permissions for a library from a user, workspace, or organization.

Download File

Download the content of a previously uploaded file from Mistral AI.

Get Agent

Tool to retrieve details of a specific Mistral AI agent by its ID.

Get Agent Version

Retrieve a specific version of an agent (Beta).

Get Batch Job

Get details and optionally inline results for a batch job.

Get campaign

Retrieve an observability campaign by its UUID.

Get campaign status

Return the execution status of an observability campaign.

Get Chat Completion Event

Retrieve one observability chat completion event by UUID.

Get Chat Completion Field Options

Get available values for a chat completion field and filter operator.

Get Connector

Get a connector by ID or name without requesting customer data or secrets.

Get Connector Authentication Methods

Get a connector's public authentication methods and required headers.

Get Conversation

Tool to retrieve details of a specific conversation.

Get Conversation History

Retrieve the full history of a conversation in Mistral AI.

Get Conversation Messages

Retrieve all messages from a Mistral AI conversation.

Get dataset

Retrieve an observability dataset by its UUID.

Get dataset import task

Retrieve the status and progress of one dataset import task.

Get dataset record

Retrieve a conversation record from an observability dataset.

Get document extracted text URL

Retrieve a signed URL to download the extracted text from a document in a Mistral AI library.

Get document signed URL

Get a signed URL to download a document from a Mistral AI library.

Get Document Status

Retrieve the processing status of a document in a Mistral AI library.

Get Document Text Content

Retrieve the extracted text content of a specific document from a Mistral AI library (Beta).

Get File Signed URL

Get a time-limited signed URL for downloading a file from Mistral AI.

List Fine Tuning Jobs

List fine-tuning jobs with optional filtering and pagination.

Get Judge

Retrieve a saved observability judge by UUID.

Get library

Retrieve detailed information about a specific library.

Get Library Document

Retrieve metadata for a specific document in a Mistral AI library.

Get Model

Tool to retrieve detailed information about a specific Mistral AI model by its ID.

Get Similar Chat Completion Events

Retrieve observability events similar to a specified event.

Get Voice

Get saved voice metadata without sample audio.

Get Voice Sample

Download a saved voice's sample as a WAV audio file.

Get Workflow

Retrieve a workflow definition and its active-worker status by name or ID.

Get Workflow Deployment

Retrieve a workflow deployment and its registered workers.

Get Workflow Execution

Retrieve the current state and result of a workflow execution.

Get Workflow Execution History

Retrieve the event history for a workflow execution.

Get Workflow Execution OpenTelemetry Trace

Retrieve raw OpenTelemetry trace data for a workflow execution.

Get Workflow Execution Trace Events

Retrieve trace events for a workflow execution.

Get Workflow Execution Trace Summary

Retrieve a hierarchical trace summary for a workflow execution.

Get Workflow Metrics

Retrieve aggregate execution, error, latency, and retry metrics for a workflow.

Get Workflow Registration

Retrieve one workflow registration and its worker availability.

Get Workflow Run

Retrieve a workflow run's status, timing, execution IDs, and result.

Get Workflow Run History

Retrieve the event history for a workflow run.

Judge Chat Completion Event

Evaluate one chat completion event with an inline judge definition.

Judge Conversation

Run a saved observability judge against a supplied conversation.

Judge dataset record

Evaluate one dataset record with a supplied non-durable judge definition.

List agent aliases

Retrieve all aliases for an agent version.

List Agents

Tool to list all configured agents (Beta).

List Agent Versions

List all versions of a specific agent.

List Batch Jobs

Tool to retrieve a list of all batch jobs with optional filtering and pagination.

List campaigns

List observability campaigns, optionally filtering them by a text query.

List campaign selected events

List chat-completion events selected by an observability campaign.

List Chat Completion Fields

List fields and operators available for observability event filtering.

List Connectors

List custom MCP connectors without returning connection secrets.

List Connector Tools

List tools exposed by an MCP connector.

List Conversations

List all created conversations (Beta).

List dataset import tasks

List import tasks associated with an observability dataset.

List dataset records

List conversation records stored in an observability dataset.

List datasets

List observability datasets, optionally filtering them by a text query.

List Files

Tool to list all files available to the user.

List Judges

List saved observability judges with filtering, search, and pagination.

List libraries

List all document libraries accessible to your organization.

List Library Documents

List all documents in a Mistral AI document library.

List library shares

List all sharing permissions for a document library.

List Models

Tool to retrieve all available Mistral AI models including base models and fine-tuned models.

List Voices

List saved voices without sample audio data.

List Workflow Deployments

List workflow worker deployments, optionally filtering to active deployments or a workflow.

List Workflow Events

List durable workflow lifecycle and task events with cursor pagination.

List Workflow Registrations

List workflow registrations with sharing, activity, archival, and cursor filters.

List Workflow Runs

List workflow runs with status, search, and cursor pagination filters.

List Workflow Schedules

List workflow schedules with timing, time zone, and execution policy details.

Query Workflow Execution

Invoke a read-only query handler on a running workflow execution.

Reprocess document

Reprocess a document in a Mistral AI library (Beta).

Restart Conversation

Tool to restart a conversation from a specific point (Beta).

Retrieve File

Retrieve metadata of a file uploaded to Mistral AI.

Search Chat Completion Events

Search observability chat completion events and return previews or only IDs.

Start Conversation

Tool to start a new conversation with a Mistral AI agent or base model.

Unarchive Fine-Tuned Model

Restore an archived fine-tuned model.

Update Agent

Tool to update an existing agent's configuration.

Update agent version

Tool to update the current version of an agent (Beta).

Update Fine-Tuned Model

Update the name or description of a fine-tuned model.

Update library

Tool to update an existing document library's properties.

Update library document

Update the metadata of a document in a Mistral AI library.

Upload File

Upload a file to Mistral AI for use in fine-tuning, batch processing, or OCR.

Upload Library Document

Upload a document to a Mistral AI library for use with RAG-enabled agents.

FRAMEWORKS

How to build Mistral ai MCP Agent with another framework

Claude Code logo

Claude Code

Use Mistral ai MCP with Claude Code

Cursor logo

Cursor

Use Mistral ai MCP with Cursor

ChatGPT logo

ChatGPT

Use Mistral ai MCP with ChatGPT

Claude Cowork logo

Claude Cowork

Use Mistral ai MCP with Claude Cowork

Antigravity logo

Antigravity

Use Mistral ai MCP with Antigravity

OpenAI Agents SDK logo

OpenAI Agents SDK

Use Mistral ai MCP with OpenAI Agents SDK

Claude Agent SDK logo

Claude Agent SDK

Use Mistral ai MCP with Claude Agent SDK

Codex logo

Codex

Use Mistral ai MCP with Codex

Grok Build logo

Grok Build

Use Mistral ai MCP with Grok Build

VS Code logo

VS Code

Use Mistral ai MCP with VS Code

OpenCode logo

OpenCode

Use Mistral ai MCP with OpenCode

OpenClaw logo

OpenClaw

Use Mistral ai MCP with OpenClaw

Hermes logo

Hermes

Use Mistral ai MCP with Hermes

Atomic Agent logo

Atomic Agent

Use Mistral ai MCP with Atomic Agent

Google ADK logo

Google ADK

Use Mistral ai MCP with Google ADK

LangChain logo

LangChain

Use Mistral ai MCP with LangChain

Mastra AI logo

Mastra AI

Use Mistral ai MCP with Mastra AI

LlamaIndex logo

LlamaIndex

Use Mistral ai MCP with LlamaIndex

CrewAI logo

CrewAI

Use Mistral ai MCP with CrewAI

Pydantic AI logo

Pydantic AI

Use Mistral ai MCP with Pydantic AI

AutoGen logo

AutoGen

Use Mistral ai MCP with AutoGen

FAQ

Frequently asked questions

With a standalone Mistral ai MCP server, the agents and LLMs can only access a fixed set of Mistral ai tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Mistral ai and many other apps based on the task at hand, all through a single MCP endpoint.

Yes, you can. Vercel AI SDK v6 fully supports MCP integration. You get structured tool calling, message history handling, and model orchestration while Tool Router takes care of discovering and serving the right Mistral ai tools.

Yes, absolutely. You can configure which Mistral ai scopes and actions are allowed when connecting your account to Composio. You can also bring your own OAuth credentials or API configuration so you keep full control over what the agent can do.

All sensitive data such as tokens, keys, and configuration is fully encrypted at rest and in transit. Composio is SOC 2 Type 2 compliant and follows strict security practices so your Mistral ai data and credentials are handled as safely as possible.

Start with Mistral ai.It takes 30 seconds.

Managed auth, hosted MCP servers, and every Mistral ai tool your agent needs.Free to start.

Start building