How to integrate Metaads MCP with Vercel AI SDK v6

Connect Vercel AI SDK v6 to Metaads MCP. Get ad performance insights for last week, create a new campaign for summer sale, and more using natural language, with authentication handled for you.

Get started for free Get a demo

Metaads

Oauth2Api Key

Metaads is Meta's official Ads API that lets you manage, analyze, and optimize your Facebook and Instagram ad campaigns. Streamline ad operations and gain deeper insights with robust automation.

70 Tools

Managed auth

Connect Metaads without auth hassles

We manage OAuth, API keys, token refresh, and scopes — you just build.

Book a call

Explore the MCP Gateway

Introduction

This guide walks you through connecting Metaads to Vercel AI SDK v6 using the Composio tool router. By the end, you'll have a working Metaads agent that can get ad performance insights for last week, create a new campaign for summer sale, pause all ads in underperforming campaigns through natural language commands.

This guide will help you understand how to give your Vercel AI SDK agent real control over a Metaads account through Composio's Metaads 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 Metaads integration
Using Composio's Tool Router to dynamically load and access Metaads 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 Metaads MCP server, and what's possible with it?

The Meta Ads MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Meta Ads account. It provides structured and secure access to your campaigns, ads, audiences, leads, performance insights, and business assets, so your agent can create campaigns, build ads, update campaigns, analyze results, and retrieve leads on your behalf.

Campaign and ad creation: Have your agent create campaigns, ad sets, ad creatives, and ads for your Meta advertising account.
Campaign updates: Direct your agent to update existing campaigns when your advertising plans or settings change.
Performance analysis: Let the agent retrieve insights for your advertising account, campaigns, ad sets, or ads to help you review results.
Audience and targeting research: Instruct your agent to create custom audiences and search available targeting options for upcoming campaigns.
Lead and asset discovery: Have your agent retrieve leads and list accessible businesses, advertising accounts, Facebook Pages, Instagram assets, campaigns, ads, and creatives.

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:

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

Step-by-step Guide

Step by step09 STEPS

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 { 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: ["metaads"],
  });

  const mcpUrl = session.mcp.url;

What's happening:

We're creating a Tool Router session that gives your agent access to Metaads 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 Metaads-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 Metaads 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 metaads, 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 Metaads 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 Metaads 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: ["metaads"],
  });

  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 metaads, 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 Metaads 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 Metaads action and event your agent gets out of the box.

Copy Ad Object

Copy a Meta Campaign, Ad Set, or Ad while explicitly selecting its resource type and supported copy options.

Create Ad

Create a new ad within an ad set using the Meta Marketing API.

Create Ad Creative

Create a new ad creative using the Meta Marketing API.

Create Ad Set

Create a new ad set within a campaign using the Meta Marketing API.

Create Ad Video

Upload an ad video asset; Meta may keep processing it after returning its ID.

Create Campaign

Create a new advertising campaign using the Meta Marketing API.

Create Custom Audience

Create a new custom audience using the Meta Marketing API.

Create Custom Conversion

Create a persistent custom-conversion definition in a Meta ad account and return its ID.

Create Lead Form

Create a persistent lead-generation form owned by a Facebook Page and return its ID.

Delete Campaign

Delete an advertising campaign using the Meta Marketing API.

Generate Ad Previews

Render previews for a provider-native creative specification in a Meta ad account and return the generated HTML bodies.

Get Ad Accounts

Tool to retrieve all ad account IDs accessible to the authenticated user from Meta Ads.

Get Ad Creative

Get Insights

Get insights data for a Meta Ads object (ad account, campaign, ad set, or ad) using the Meta Marketing API.

Get Meta Object

Tool to retrieve data for any Meta Marketing API object by its ID.

Get Page Accounts

Retrieve pages managed by the authenticated user, including page access tokens when requested.

Get Pixel Event Stats

Read recent Pixel event-count and processing statistics, optionally restricted to server events and one event name, to verify aggregate ingestion.

Get User

Tool to retrieve information about the authenticated user from Meta (Facebook) Graph API.

List Activities

List activity history for a Meta ad account or ad set.

List Ad Account Pixels

List the Pixels and consolidated datasets associated with an ad account so a usable data-source ID can be selected for Conversions API events.

List Ad Creatives

Tool to list all ad creatives under an ad account.

List Ad Images

List one page of image assets in a Meta ad account, optionally restricted to exact image hashes.

List Ad Network Analytics

Tool to retrieve ad network analytics for a Meta Business using the Meta Marketing API.

List Ad Network Analytics Results

Tool to retrieve ad network analytics results for Facebook Audience Network.

List Ads

Tool to list all ads under an ad account using the Meta Marketing API.

List Ad Videos

List one page of video assets in a Meta ad account, optionally filtered by title, dimensions, duration, or aspect ratio.

List Agencies

Tool to retrieve agencies associated with a Meta Business or Ad Account.

List Assigned Pages

Tool to retrieve Facebook Pages assigned to a business user via Meta Marketing API.

List Assigned Users

Tool to list users assigned to a Facebook Page or Ad Account within a specific business context.

List Business Ad Accounts

Tool to retrieve all ad accounts owned by a specific Business Manager.

List Businesses

Tool to discover every Business Manager the authenticated Meta principal can access, via `GET /me/businesses`.

List Business Invoices

Tool to retrieve business invoices from Meta Marketing API.

List Campaigns

List one page of campaigns belonging to a Meta ad account.

List Client Ad Accounts

Tool to list all client ad accounts accessible to a business from Meta Ads.

List Client Apps

Tool to retrieve client apps associated with a Meta Business using the Marketing API.

List Client Instagram Assets

Tool to retrieve Instagram assets that are shared with a business as a client using the Meta Marketing API.

List Client Offsite Signal Container Business Objects

Tool to retrieve client offsite signal container business objects for a business from Meta Marketing API.

List Client Pages

Tool to retrieve client pages associated with a Meta business.

List Clients

Tool to retrieve client businesses associated with a Meta Business Manager.

List Collaborative Ads Collaboration Requests

Tool to retrieve collaborative ads collaboration requests for a Meta Business using the Marketing API.

List Collaborative Ads Suggested Partners

Tool to retrieve collaborative ads suggested partners for a business from Meta Marketing API.

List Comments

Return one page from the comments edge of a Page-owned post, photo, video, comment, or reply.

List Custom Audiences

List one page of custom audiences in a Meta ad account, with field projection and optional business, primary-audience, or Pixel constraints.

List Custom Conversions

List one page of custom-conversion definitions in a Meta ad account, with field projection and opaque cursor pagination.

List Initiated Audience Sharing Requests

Tool to retrieve initiated audience sharing requests for a business using the Meta Marketing API.

List Leads

List one page of leads belonging to a Meta Ad or Leadgen Form.

List Managed Partner Ads Funding Source Details

Tool to retrieve managed partner ads funding source details for a Meta Business.

List Owned Apps

Tool to retrieve apps owned by a business from Meta Marketing API.

List Owned Businesses

Tool to retrieve businesses owned by a parent Business Manager from Meta Marketing API.

List Owned Instagram Assets

Tool to list Instagram accounts/assets owned by a Meta Business.

List Owned Offsite Signal Container Business Objects

Tool to retrieve owned offsite signal container business objects for a business from Meta Marketing API.

List Owned Pages

Tool to retrieve Pages owned by a Business Manager from Meta Marketing API.

List Pending Client Ad Accounts

Tool to retrieve pending client ad account access requests for a Business Manager from Meta Ads.

List Pending Client Apps

Tool to retrieve pending client apps for a Meta Business.

List Pending Client Pages

Tool to retrieve pending client pages for a Business Manager using the Meta Marketing API.

List Pending Shared Offsite Signal Container Business Objects

Tool to retrieve pending shared offsite signal container business objects from Meta Marketing API.

List Pending Owned Ad Accounts

Tool to retrieve pending owned ad accounts for a Business Manager account from Meta Marketing API.

List Pending Owned Pages

Tool to retrieve Pages with pending ownership status for a Business Manager from Meta Marketing API.

List Pending Users

Tool to retrieve pending users from a Business Manager in Meta Marketing API.

List Received Audience Sharing Requests

Tool to retrieve all received audience sharing requests for a business using the Meta Marketing API.

List Subscribed Apps

List one page of applications subscribed to a Meta ad account.

List System Users

Tool to retrieve system users for a Meta Business Manager account.

List Targeting Search

Tool to search for targeting options in Meta Ads Marketing API.

List Video Tags

Return one page of TaggableSubject entries from a Video's `/tags` edge.

List Video Thumbnails

List one page of available thumbnail variants for a Meta video.

Read Ad Sets

Retrieve ad sets from a Meta ad account using the Marketing API.

Send Conversion Events

Irreversibly submit 1-1,000 validated server conversion events to a Meta Pixel or dataset.

Update Ad Creative

Update Campaign

Update an existing advertising campaign using the Meta Marketing API.

Upload Ad Image

Upload an image for use in Meta ad creatives using the Marketing API.

FRAMEWORKS

How to build Metaads MCP Agent with another framework

Claude Code

Use Metaads MCP with Claude Code

Cursor

Use Metaads MCP with Cursor

ChatGPT

Use Metaads MCP with ChatGPT

Claude Cowork

Use Metaads MCP with Claude Cowork

Antigravity

Use Metaads MCP with Antigravity

OpenAI Agents SDK

Use Metaads MCP with OpenAI Agents SDK

Claude Agent SDK

Use Metaads MCP with Claude Agent SDK

Grok Bot

Use Metaads MCP with Grok Bot

Muse

Use Metaads MCP with Muse

Codex

Use Metaads MCP with Codex

Kimi Code

Use Metaads MCP with Kimi Code

Grok Build

Use Metaads MCP with Grok Build

VS Code

Use Metaads MCP with VS Code

OpenCode

Use Metaads MCP with OpenCode

OpenClaw

Use Metaads MCP with OpenClaw

Hermes

Use Metaads MCP with Hermes

Atomic Agent

Use Metaads MCP with Atomic Agent

CLI

Use Metaads MCP with CLI

Google ADK

Use Metaads MCP with Google ADK

LangChain

Use Metaads MCP with LangChain

Mastra AI

Use Metaads MCP with Mastra AI

LlamaIndex

Use Metaads MCP with LlamaIndex

CrewAI

Use Metaads MCP with CrewAI

Pydantic AI

Use Metaads MCP with Pydantic AI

AutoGen

Use Metaads MCP with AutoGen

MORE TOOLKITS

Explore Other Toolkits

Toolkit marketplace

Adrapid

Api Key

Adrapid is a platform for rapid creation of digital marketing visuals using templates. It streamlines design workflows for banners, images, and HTML5 content with automation.

AdRoll

Oauth2

AdRoll is a digital marketing platform for managing advertising campaigns, audiences, and account-based marketing. It helps marketers retarget customers across channels and measure performance.

Adtraction

Api Key

Adtraction is an affiliate marketing platform for partner programs, tracking links, transactions, commissions, statistics, and payments. It helps advertisers and publishers manage performance marketing with clear reporting and payout workflows.

Adyntel

Api Key

Adyntel is an API that retrieves LinkedIn ads for any company using a domain or LinkedIn Page ID. Easily access competitive ad intelligence to power your marketing workflows.

FAQ

Frequently asked questions

With a standalone Metaads MCP server, the agents and LLMs can only access a fixed set of Metaads tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Metaads 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 Metaads tools.

Yes, absolutely. You can configure which Metaads 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 Metaads data and credentials are handled as safely as possible.

Start with Metaads.It takes 30 seconds.

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

Start building

How to integrate Metaads MCP with Vercel AI SDK v6

Connect Metaads without auth hassles

Introduction

Also integrate Metaads with

TL;DR

What is Vercel AI SDK?

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

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

What is Composio SDK?

How the Composio SDK works

Step-by-step Guide

Prerequisites

Getting API Keys for OpenAI and Composio

Install required dependencies

Set up environment variables

Import required modules and validate environment

Create Tool Router session and initialize MCP client

Connect to MCP server and retrieve tools

Initialize conversation and CLI interface

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

Complete Code

Conclusion

Supported Tools

How to build Metaads MCP Agent with another framework

Claude Code

Cursor

ChatGPT

Claude Cowork

Antigravity

OpenAI Agents SDK

Claude Agent SDK

Grok Bot

Muse

Codex

Kimi Code

Grok Build

VS Code

OpenCode

OpenClaw

Hermes

Atomic Agent

CLI

Google ADK

LangChain

Mastra AI

LlamaIndex

CrewAI

Pydantic AI

AutoGen

Explore Other Toolkits

Adrapid

AdRoll

Adtraction

Adyntel

Frequently asked questions

What are the differences in Tool Router MCP and Metaads MCP?+

Can I use Tool Router MCP with Vercel AI SDK v6?+

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

How safe is my data with Composio Tool Router?+

Start with Metaads.It takes 30 seconds.