How to integrate Mem0 MCP with Vercel AI SDK

BOOK A CALLGET STARTED FOR FREE
Framework Integration Gradient
Mem0 Logo
Vercel AI SDK Logo
divider

Introduction

This guide walks you through connecting Mem0 to Vercel AI SDK using the Composio tool router. By the end, you'll have a working Mem0 agent that can store meeting notes from today's call, export all project memories as csv, add new user to our team space, search recent notes mentioning quarterly goals through natural language commands.

This guide will help you understand how to give your Vercel AI SDK agent real control over a Mem0 account through Composio's Mem0 MCP server.

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

TL;DR

Here's what you'll learn:
  • How to set up and configure a Vercel AI SDK agent with Mem0 integration
  • Using Composio's Tool Router to dynamically load and access Mem0 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
  • Step Counting: Control multi-step tool execution
  • OpenAI Provider: Native integration with OpenAI models

What is the Mem0 MCP server, and what's possible with it?

The Mem0 MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Mem0 account. It provides structured and secure access to your notes, projects, and organizational knowledge, so your agent can perform actions like searching memories, managing users, adding content, and orchestrating agent runs on your behalf.

  • AI-powered memory search and recall: Let your agent search and retrieve existing memory entries using advanced filters and pagination to surface just the right note or piece of information.
  • Automated content and note creation: Have your agent store new memory records from conversations, meetings, or tasks—ensuring nothing slips through the cracks.
  • Collaboration and organization management: Direct your agent to add members to projects or organizations, assign roles, and keep team structures up to date.
  • Agent and application orchestration: Enable your agent to create new AI agents, initiate agent runs, and manage applications for custom workflows and automation.
  • Structured knowledge export and reporting: Ask your agent to initiate export jobs with specific schemas and filters, so you can back up or analyze your stored knowledge on demand.

Supported Tools & Triggers

Tools
Add member to projectAdds an existing user to a project (identified by `project id` within organization `org id`), assigning a valid system role.
Add new memory recordsStores new memory records from a list of messages, optionally inferring structured content; requires association via `agent id`, `user id`, `app id`, or `run id`.
Add organization memberAdds a new member, who must be a registered user, to an organization, assigning them a specific role.
Create a new agentCreates a new agent with a unique `agent id` and an optional `name`; additional metadata may be assigned by the system.
Create a new agent runCreates a new agent run in the mem0.
Create a new applicationCreates a new application, allowing metadata to be passed in the request body (not an explicit field in this action's request model); ensure `app id` is unique to avoid potential errors or unintended updates.
Create a new organization entryCreates a new organization entry using the provided name and returns its details.
Create a new userCreates a new user with the specified unique `user id` and supports associating `metadata` (not part of the request schema fields).
Create an export job with schemaInitiates an asynchronous job to export memories, structured by a schema provided in the request body and allowing optional filters.
Create memory entryLists/searches existing memory entries with filtering and pagination; critically, this action retrieves memories and does *not* create new ones, despite its name.
Create projectCreates a new project with a given name within an organization that must already exist.
Delete an organizationPermanently deletes an existing organization identified by its unique id.
Delete memory by idPermanently deletes a specific memory by its unique id; ensure the `memory id` exists as this operation is irreversible.
Delete entity by type and idCall to permanently and irreversibly hard-delete an existing entity (user, agent, app, or run) and all its associated data, using its type and id.
Delete memoriesDeletes memories matching specified filter criteria; omitting all filters may result in deleting all memories.
Delete memory batch with uuidsDeletes a batch of up to 1000 existing memories, identified by their uuids, in a single api call.
Delete projectPermanently deletes a specific project and all its associated data from an organization; this action cannot be undone and requires the project to exist within the specified organization.
Delete project memberRemoves an existing member, specified by username, from a project, immediately revoking their project-specific access; the user is not removed from the organization.
Export data based on filtersRetrieves memory export data, optionally filtered by various identifiers (e.
List organizationsRetrieves a summary list of organizations for administrative oversight; returns summary data (names, ids), not exhaustive details, despite 'detailed' in the name.
Fetch details of a specific organizationFetches comprehensive details for an organization using its `org id`; the `org id` must be valid and for an existing organization.
Get list of entity filtersRetrieves predefined filter definitions for entities (e.
Get entity by idFetches detailed information for an existing entity (user, agent, app, or run) identified by its type and unique id.
Get organization membersFetches a list of members for a specified, existing organization.
Get project detailsFetches comprehensive details for a specified project within an organization.
Get project membersRetrieves all members for a specified project within an organization.
Get projectsRetrieves all projects for a given organization `org id` to which the caller has access.
Get user memory statsRetrieves a summary of the authenticated user's memory activity, including total memories created, search events, and add events.
List entitiesRetrieves a list of entities, optionally filtered by organization or project (prefer `org id`/`project id` over deprecated `org name`/`project name`), noting results may be summaries and subject to limits.
Perform semantic search on memoriesSearches memories semantically using a natural language query (required if `only metadata based search` is false) and/or metadata filters.
Remove a member from the organizationRemoves a member, specified by their username, from an existing organization of which they are currently a member.
Retrieve all events for the currently logged in userRetrieves a paginated list of events for the authenticated user, filterable and paginable via url query parameters.
Retrieve entity-specific memoriesRetrieves all memories (e.
Retrieve list of memory eventsRetrieves a chronological list of all memory events (e.
Retrieve memory by idRetrieves a complete memory entry by its unique identifier; `memory id` must be valid and for an existing memory.
Retrieve memory history by idRetrieves the complete version history for an existing memory, using its unique `memory id`, to inspect its evolution or audit changes.
Retrieve memory listRetrieves a list of memories, supporting pagination and diverse filtering (e.
Search memories with filtersSemantically searches memories using a natural language query and mandatory structured filters, offering options to rerank results and select specific fields; any provided `org id` or `project id` must reference a valid existing entity.
Update memory batch with uuidUpdates text for up to 1000 memories in a single batch, using their uuids.
Update memory text contentUpdates the text content of an existing memory, identified by its `memory id`.
Update organization member roleUpdates the role of an existing member to a new valid role within an existing organization.
Update projectUpdates a project by `project id` within an `org id`, modifying only provided fields (name, description, custom instructions, custom categories); list fields are fully replaced (cleared by `[]`), other omitted/null fields remain unchanged.
Update project member roleUpdates the role of a specific member within a designated project, ensuring the new role is valid and recognized by the system.

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

What is Tool Router?

Composio's Tool Router 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 Tool Router

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

How the Tool Router works

The Tool Router 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

Prerequisites

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

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.

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

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

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 { experimental_createMCPClient as 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

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: ["mem0"],
  });

  const mcpUrl = session.mcp.url;
What's happening:
  • We're creating a Tool Router session that gives your agent access to Mem0 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 Mem0-related tools through the MCP protocol

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 Mem0 tools that the agent can use

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 mem0, 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

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 Mem0 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 Mem0 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 { experimental_createMCPClient as 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: ["mem0"],
  });

  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 mem0, 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 Mem0 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.

How to build Mem0 MCP Agent with another framework

OpenAI Agents SDK

OpenAI Agents SDK

Use Mem0 MCP with OpenAI Agents SDK

Claude Agent SDK

Claude Agent SDK

Use Mem0 MCP with Claude Agent SDK

Google ADK

Google ADK

Use Mem0 MCP with Google ADK

LangChain

LangChain

Use Mem0 MCP with LangChain

Mastra AI

Mastra AI

Use Mem0 MCP with Mastra AI

LlamaIndex

LlamaIndex

Use Mem0 MCP with LlamaIndex

CrewAI

CrewAI

Use Mem0 MCP with CrewAI

FAQ

What are the differences in Tool Router MCP and Mem0 MCP?

With a standalone Mem0 MCP server, the agents and LLMs can only access a fixed set of Mem0 tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Mem0 and many other apps based on the task at hand, all through a single MCP endpoint.

Can I use Tool Router MCP with Vercel AI SDK?

Yes, you can. Vercel AI SDK 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 Mem0 tools.

Can I manage the permissions and scopes for Mem0 while using Tool Router?

Yes, absolutely. You can configure which Mem0 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.

How safe is my data with Composio Tool Router?

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 Mem0 data and credentials are handled as safely as possible.

Used by agents from

Context
ASU
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai
Context
ASU
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai
Context
ASU
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai

Never worry about agent reliability

We handle tool reliability, observability, and security so you never have to second-guess an agent action.

Book a call↗Get started for free