How to integrate Fireberry MCP with Vercel AI SDK v6

Trusted by
AWS
Glean
Zoom
Airtable
GET STARTED FOR FREEGET A DEMO

30 min · no commitment · see it on your stack

Fireberry logo
Vercel AI SDK logo
divider

Introduction

This guide walks you through connecting Fireberry to Vercel AI SDK v6 using the Composio tool router. By the end, you'll have a working Fireberry agent that can add new lead to contacts table, list all open deals in pipeline, fetch picklist options for deal stage through natural language commands.

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

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

Also integrate Fireberry with

TL;DR

Here's what you'll learn:
  • How to set up and configure a Vercel AI SDK agent with Fireberry integration
  • Using Composio's Tool Router to dynamically load and access Fireberry 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 Fireberry MCP server, and what's possible with it?

The Fireberry MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Fireberry CRM account. It provides structured and secure access to your CRM data, so your agent can perform actions like creating records, querying customer information, and managing picklists on your behalf.

  • Automated record creation: Let your agent swiftly add new entries to any Fireberry table, such as contacts, leads, or deals, using structured data you provide.
  • Smart CRM data retrieval: Ask your agent to query records with powerful filtering, sorting, and pagination—perfect for finding the exact customer or deal you need.
  • Picklist value management: Effortlessly fetch all available options for any picklist (dropdown) field, making data entry and workflow automation simpler and error-free.
  • Custom module support: Enable your agent to work with any Fireberry module, so you can handle specialized business processes or custom workflows.

Supported Tools & Triggers

Tools
Create a CompetitorTool to create a new competitor in Fireberry.
Create a Fireberry contactTool to create a new contact in Fireberry CRM.
Create an Activity LogCreate a new Activity Log record in Fireberry.
Create a date fieldTool to create a new date field in a Fireberry object/table.
Create a Fireberry lookup fieldTool to create a lookup field in Fireberry CRM.
Create a new Fireberry accountTool to create a new account in Fireberry CRM.
Create an ArticleTool to create a new article in Fireberry.
Create a Fireberry assetCreates a new asset (account product) in Fireberry.
Create an OpportunityTool to create a new opportunity in Fireberry CRM.
Create an Order ItemTool to create a new order item in Fireberry CRM.
Create a noteCreate a new note record in Fireberry.
Create a phone callCreates a new phone call record in Fireberry's call log.
Create a Fireberry productTool to create a new product in Fireberry.
Create a ProjectTool to create a new project in Fireberry CRM.
Create a TaskTool to create a new task in Fireberry CRM.
Create a TicketTool to create a new ticket (case) in Fireberry CRM.
Create a URL fieldTool to create a new URL field in a Fireberry object/table.
Create a CampaignTool to create a new campaign in Fireberry.
Create a CRM OrderTool to create a new CRM Order in Fireberry.
Upload file to Fireberry recordTool to upload a file to a specific record in Fireberry.
Create a Fireberry meetingTool to create a new meeting (activity) in Fireberry.
Create a new Fireberry recordCreates a new record in a specified Fireberry table/module.
Delete an AccountTool to delete an account from Fireberry using its GUID.
Delete a CompetitorTool to delete a competitor in Fireberry by its GUID.
Delete a contractTool to delete a contract in Fireberry by its GUID.
Delete an Activity LogDelete an activity log by its GUID.
Delete a custom fieldTool to delete a custom field from a Fireberry object/table.
Delete an ArticleTool to delete an article from Fireberry by its GUID.
Delete a Fireberry assetDelete an asset from Fireberry by its unique identifier.
Delete an OpportunityTool to delete an opportunity in Fireberry by its GUID.
Delete an Order ItemTool to delete an order item in Fireberry by its GUID.
Delete a NoteTool to delete a note from Fireberry using its GUID.
Delete a Phone CallTool to delete a phone call record from Fireberry using its GUID.
Delete a projectTool to delete a project from Fireberry using its GUID.
Delete a taskTool to delete a task in Fireberry by its GUID.
Delete a ticketTool to delete a ticket (case) from Fireberry using its GUID.
Delete a Business UnitTool to delete a Business Unit in Fireberry using its GUID.
Delete a Fireberry campaignTool to delete a campaign from Fireberry.
Delete a ContactTool to delete a contact from Fireberry using its GUID.
Delete a CRM UserTool to delete a CRM user from Fireberry using its GUID.
Delete a MeetingTool to delete a meeting activity from Fireberry by its GUID.
Delete a productTool to delete a product in Fireberry by its GUID.
Get an AccountTool to retrieve a specific account record by its GUID.
Get a CRM OrderTool to retrieve a specific CRM Order from Fireberry by its GUID.
Get All AccountsTool to retrieve all accounts from Fireberry CRM with pagination support.
Get All Activity Logs (v2)Tool to retrieve all activity logs from Fireberry using v2 API endpoint with pagination support.
Get all articles from FireberryTool to retrieve all articles from Fireberry with pagination support.
Get All AssetsTool to retrieve all assets (account products) from Fireberry with pagination support.
Get All Business Units (v2)Tool to retrieve all business units from Fireberry using v2 API endpoint with pagination support.
Get All CampaignsTool to retrieve all campaigns from Fireberry with pagination support.
Get All Competitors (v2)Tool to retrieve all competitors from Fireberry with pagination support.
Get All ContactsTool to retrieve all contacts from Fireberry with pagination support.
Get All ContractsTool to retrieve all contracts from Fireberry with pagination support.
Get All Custom Object RecordsTool to retrieve all records from a specified custom object in Fireberry with pagination support.
Get All MeetingsTool to retrieve all meetings (activities) from Fireberry CRM with pagination support.
Get All Notes (Detailed)Tool to retrieve all notes from Fireberry with detailed field schema and pagination support.
Get All ObjectsTool to retrieve all object type metadata from Fireberry.
Get All Order ItemsTool to retrieve all order items from Fireberry with pagination support.
Get All OrdersTool to retrieve all orders from Fireberry with pagination support.
Get All Phone CallsTool to retrieve all phone call records from Fireberry with pagination support.
Get All Projects (v2)Tool to retrieve all projects from Fireberry with pagination support.
Get All TasksTool to retrieve all tasks from Fireberry with pagination support.
Get All TicketsTool to retrieve all ticket records (cases) from Fireberry with pagination support.
Get All UsersTool to retrieve all CRM users from Fireberry with pagination support.
Get a MeetingTool to retrieve a specific meeting/activity record by its unique identifier (GUID).
Get an Activity LogTool to retrieve a specific activity log record from Fireberry by its GUID.
Get an ArticleTool to retrieve a specific article from Fireberry by its GUID.
Get an AssetTool to retrieve a specific asset record by its GUID.
Get an ObjectTool to retrieve metadata for a specific object by its ID.
Get an Object's FieldsTool to retrieve metadata about fields for a specific object type in Fireberry.
Get an OpportunityTool to retrieve a specific opportunity record by its GUID.
Get an Order ItemTool to retrieve a specific order item record by its GUID.
Get a NoteTool to retrieve a specific note record by its GUID.
Get a phone call recordTool to retrieve a specific phone call record from Fireberry by its GUID.
Get a ProductTool to retrieve a specific product record by its GUID.
Get a ProjectTool to retrieve a specific project from Fireberry by its GUID.
Get a TaskTool to retrieve a specific task record by its GUID.
Get a TicketTool to retrieve a specific ticket (case) record by its GUID.
Get Campaign by IDTool to retrieve a single campaign by its GUID.
Get a CompetitorTool to retrieve a specific competitor record by its GUID.
Get a ContactTool to retrieve a specific contact record by its GUID.
Get Custom Object RecordTool to retrieve a specific custom object record by its GUID and object code.
Get Field DetailsTool to retrieve detailed metadata for a specific field in a Fireberry object/table.
Get Object Field ValuesTool to retrieve picklist field values from the metadata endpoint.
Get Items for an OrderTool to retrieve all items for a specific order from Fireberry.
Get Picklist Field ValuesTool to retrieve picklist field values from Fireberry metadata API.
Get Picklist ValuesTool to retrieve all possible picklist (dropdown) values for a specific field by querying records and extracting unique values.
Get Related RecordsTool to retrieve related records for a specific object in Fireberry.
Get Fireberry Task by IDTool to retrieve a single task record by its unique ID (GUID).
Get a Fireberry user by IDTool to retrieve a single user by their unique ID from Fireberry.
List All OpportunitiesTool to retrieve all opportunities from Fireberry CRM with pagination support.
List All ProductsTool to retrieve all products from Fireberry CRM with pagination support.
Fireberry: Query RecordsQuery and retrieve records from a Fireberry module with optional filtering, sorting, and pagination.
Query Fireberry records with filtersQuery records in any Fireberry object with advanced filtering, sorting, and pagination.
Update a Business UnitTool to update an existing business unit in Fireberry.
Update Fireberry AccountUpdates an existing account record in Fireberry with new field values.
Update a Fireberry CompetitorUpdates an existing competitor record in Fireberry by GUID.
Update a Fireberry contactTool to update an existing contact in Fireberry CRM.
Update a ContractTool to update an existing contract in Fireberry.
Update an Activity LogUpdate an existing Activity Log record in Fireberry.
Update a Date FieldTool to update a date field configuration in Fireberry.
Update a Date & Time FieldTool to update a Date & Time field's properties in Fireberry.
Update a Formula FieldTool to update a formula field in Fireberry CRM.
Update an HTML FieldTool to update an HTML field configuration in Fireberry.
Update a Fireberry MeetingTool to update an existing meeting (activity) in Fireberry.
Update a Fireberry articleUpdates an existing article in Fireberry.
Update an AssetUpdate an existing asset (accountproduct) in Fireberry.
Update an Email Address FieldTool to update the configuration of an email address field in Fireberry.
Update an OpportunityTool to update an existing opportunity in Fireberry CRM.
Update an Order ItemTool to update an existing order item in Fireberry.
Update a Number FieldTool to update a number field configuration in Fireberry.
Update a Phone Number FieldTool to update a phone number field configuration in Fireberry.
Update a ProductTool to update an existing product in Fireberry.
Update a ProjectTool to update an existing project in Fireberry CRM.
Update a Text Area FieldTool to update a Text Area field's properties in Fireberry.
Update a Text FieldTool to update a text field configuration in Fireberry.
Update a TicketTool to update an existing ticket (case) in Fireberry.
Update a URL FieldTool to update a URL field configuration in Fireberry.
Update a UserTool to update an existing user in Fireberry CRM.
Update a Fireberry CampaignTool to update an existing campaign in Fireberry by its GUID.
Update a CRM OrderTool to update an existing CRM order in Fireberry.
Update a phone call recordTool to update an existing phone call record in Fireberry.
Update a Task (V2)Tool to update an existing task using Fireberry v2 API.

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

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: ["fireberry"],
  });

  const mcpUrl = session.mcp.url;
What's happening:
  • We're creating a Tool Router session that gives your agent access to Fireberry 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 Fireberry-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 Fireberry 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 fireberry, 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 Fireberry 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 Fireberry 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: ["fireberry"],
  });

  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 fireberry, 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 Fireberry 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 Fireberry MCP Agent with another framework

OpenAI Agents SDK logo

OpenAI Agents SDK

Use Fireberry MCP with OpenAI Agents SDK

Claude Agent SDK logo

Claude Agent SDK

Use Fireberry MCP with Claude Agent SDK

Claude Code logo

Claude Code

Use Fireberry MCP with Claude Code

Claude Cowork logo

Claude Cowork

Use Fireberry MCP with Claude Cowork

Codex logo

Codex

Use Fireberry MCP with Codex

OpenClaw logo

OpenClaw

Use Fireberry MCP with OpenClaw

Hermes logo

Hermes

Use Fireberry MCP with Hermes

CLI logo

CLI

Use Fireberry MCP with CLI

Google ADK logo

Google ADK

Use Fireberry MCP with Google ADK

LangChain logo

LangChain

Use Fireberry MCP with LangChain

Mastra AI logo

Mastra AI

Use Fireberry MCP with Mastra AI

LlamaIndex logo

LlamaIndex

Use Fireberry MCP with LlamaIndex

CrewAI logo

CrewAI

Use Fireberry MCP with CrewAI

FAQ

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

With a standalone Fireberry MCP server, the agents and LLMs can only access a fixed set of Fireberry tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Fireberry 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 v6?

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 Fireberry tools.

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

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

Used by agents from

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

Get started for freeGet a demo↗