# How to integrate Bart MCP with Vercel AI SDK v6 ```json { "title": "How to integrate Bart MCP with Vercel AI SDK v6", "toolkit": "Bart", "toolkit_slug": "bart", "framework": "Vercel AI SDK", "framework_slug": "ai-sdk", "url": "https://composio.dev/toolkits/bart/framework/ai-sdk", "markdown_url": "https://composio.dev/toolkits/bart/framework/ai-sdk.md", "updated_at": "2026-05-12T10:02:26.975Z" } ``` ## Introduction This guide walks you through connecting Bart to Vercel AI SDK v6 using the Composio tool router. By the end, you'll have a working Bart agent that can find next departures from embarcadero station, get real-time trip updates for richmond line, check current bart service advisories and alerts through natural language commands. This guide will help you understand how to give your Vercel AI SDK agent real control over a Bart account through Composio's Bart MCP server. Before we dive in, let's take a quick look at the key ideas and tools involved. ## Also integrate Bart with - [OpenAI Agents SDK](https://composio.dev/toolkits/bart/framework/open-ai-agents-sdk) - [Claude Agent SDK](https://composio.dev/toolkits/bart/framework/claude-agents-sdk) - [Claude Code](https://composio.dev/toolkits/bart/framework/claude-code) - [Claude Cowork](https://composio.dev/toolkits/bart/framework/claude-cowork) - [Codex](https://composio.dev/toolkits/bart/framework/codex) - [OpenClaw](https://composio.dev/toolkits/bart/framework/openclaw) - [Hermes](https://composio.dev/toolkits/bart/framework/hermes-agent) - [CLI](https://composio.dev/toolkits/bart/framework/cli) - [Google ADK](https://composio.dev/toolkits/bart/framework/google-adk) - [LangChain](https://composio.dev/toolkits/bart/framework/langchain) - [Mastra AI](https://composio.dev/toolkits/bart/framework/mastra-ai) - [LlamaIndex](https://composio.dev/toolkits/bart/framework/llama-index) - [CrewAI](https://composio.dev/toolkits/bart/framework/crew-ai) ## TL;DR Here's what you'll learn: - How to set up and configure a Vercel AI SDK agent with Bart integration - Using Composio's Tool Router to dynamically load and access Bart 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 Bart MCP server, and what's possible with it? The Bart MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to BART's public transit data. It provides structured and secure access to real-time schedules, route information, station details, and service advisories, so your agent can plan trips, fetch live updates, check advisories, and explore routes for you. - Trip planning with live schedules: Instantly retrieve train arrival or departure times and help users plan journeys between any BART stations based on the latest schedule data. - Live service advisories and alerts: Keep travelers informed by fetching up-to-date system-wide or station-specific service advisories, ensuring users know about delays or disruptions before they travel. - Route and station discovery: Access detailed information about BART routes and stations, including amenities and configuration, so your agent can answer travel questions or recommend stations. - Real-time trip and schedule updates: Get the latest trip updates and schedule changes in real time, allowing users to adapt plans quickly if there are changes or issues along their route. - Access to static and GTFS feeds: Download the latest BART GTFS (General Transit Feed Specification) data for offline schedule planning, analysis, or integration with third-party transit tools. ## Supported Tools | Tool slug | Name | Description | |---|---|---| | `BART_BART_GET_API_VERSION` | Get BART API Version | Get the current version of the BART API. This action retrieves version information for the BART (Bay Area Rapid Transit) API, including the current API version number, copyright information, and license details. This is useful for verifying API compatibility and ensuring you're working with the expected API version. The BART API is currently at version 3.10 and supports both XML and JSON output formats. Use this action to confirm which version of the API you're interfacing with and to access licensing information. | | `BART_GET_ELEVATOR_STATUS` | Get Elevator Status | Tool to fetch current elevator status across all BART stations. Use when you need real-time elevator availability information for accessibility planning or route guidance. | | `BART_GET_ESTIMATED_DEPARTURES` | Get Estimated Departures | Tool to get real-time estimated departure times for a specified BART station. Returns live train departure predictions including minutes until departure, platform assignments, train lengths, line colors, bicycle accommodation, and delay information. Use this when you need current departure times for planning trips or checking train status. | | `BART_GET_FARE` | Get BART Fare | Get fare information between two BART stations including Clipper and cash prices. Returns multiple fare types (Clipper, cash, senior/disabled, youth, Clipper START) with their respective prices. Use this when you need to find out how much a BART trip costs between two stations. | | `BART_GET_GTFS_ALERTS` | Get GTFS-RT Service Alerts | Tool to fetch GTFS-RT service alerts in protobuf format for integration with GTFS static feed. Use when you need real-time service advisories, disruptions, or alert information. | | `BART_GET_GTFS_RT_TRIP_UPDATES` | Get GTFS-RT Trip Updates | Tool to fetch real-time trip updates in GTFS-Realtime format. Use when you need the latest live trip information as raw protobuf. | | `BART_GET_GTFS_STATIC_SCHEDULE_FEED` | Download GTFS Static Schedule Feed | Downloads the BART static GTFS (General Transit Feed Specification) schedule feed as a ZIP archive. The GTFS feed contains comprehensive transit data including stations, routes, trip schedules, fares, and service calendars in standardized CSV format. Use this to access complete BART schedule information for route planning, analysis, or integration with transit applications. | | `BART_BART_GET_ROUTE_INFO` | Get Route Info | Tool to fetch detailed information about a specific BART route. Use when you know the route number (1โ€“12) or need all routes configuration. Call after confirming the route ID. | | `BART_GET_ROUTE_SCHEDULE` | Get Route Schedule | Tool to get detailed schedule information for a specific BART route showing all trains and their stops. Use when you need to see the complete schedule for a route including departure times, station stops, bike policies, and passenger load indicators. Call this after determining the specific route number (1-12). | | `BART_GET_SCHEDULE_ARRIVE` | Get BART Schedule Arrive | Tool to retrieve schedule information based on a specified arrival time. Use when planning trips arriving by a given time. | | `BART_GET_SCHEDULE_DEPART` | Get BART Schedule Depart | Get BART train schedules departing from an origin station to a destination station at a specified time. Returns multiple trip options with departure/arrival times, fares (Clipper, cash, senior/disabled, youth), transfer details, train information, and platform numbers. Use this when you need to plan BART trips with specific departure times or when users ask about train schedules between two stations. | | `BART_GET_SERVICE_ADVISORIES` | Get Service Advisories | Tool to fetch current BART service advisories. Use when you need up-to-date system-wide or station-level alerts before presenting or planning transit routes. | | `BART_GET_STATION_ACCESS` | Get Station Access | Get comprehensive station access information including parking, transit, bike facilities, and lockers. Returns detailed access information for a specific BART station including: entering/exiting instructions, parking availability and lot capacity, bike parking and bike station details, locker availability, car-sharing options, nearby destinations, and connected transit services. Use this when you need to help users understand how to access a BART station or what facilities are available. | | `BART_GET_STATION_INFO` | Get Station Info | Get detailed information for a specific BART station by its abbreviation code. Returns comprehensive station details including: name, location (address, city, county, coordinates), routes serving the station (northbound/southbound), platform information, nearby amenities (food, shopping, attractions), and general station description. Use this when you need detailed information about a specific BART station and you already have its 4-letter abbreviation code (e.g., 'EMBR' for Embarcadero, 'MONT' for Montgomery Street, '12TH' for 12th Street Oakland). | | `BART_BART_GET_STATIONS` | Get BART Stations | Get a list of all BART stations with their complete information. This action retrieves information about all BART (Bay Area Rapid Transit) stations including station names, abbreviation codes, geographic coordinates (latitude/longitude), and full addresses. This is useful for finding station locations, getting station codes for other API calls, or building station lookup tools. | | `BART_GET_STATION_SCHEDULE` | Get Station Schedule | Get detailed scheduled departure information for a specific BART station. Returns all trains departing from the station including route line, destination, departure time, bike allowance, crowding level, and platform number. Use this when you need to see all departures from a specific station. | | `BART_GET_TRAIN_COUNT` | Get Train Count | Tool to fetch current count of trains active in the BART system. Use when you need real-time information about how many trains are currently operating. | | `BART_LIST_ROUTES` | List BART Routes | Tool to get a list of all current BART routes/lines with basic information. Use when you need to see all available routes, their colors, directions, or route numbers. | ## Supported Triggers None listed. ## Creating MCP Server - Stand-alone vs Composio SDK The Bart MCP server is an implementation of the Model Context Protocol that connects your AI agent to Bart. It provides structured and secure access so your agent can perform Bart operations on your behalf through a secure, permission-based interface. With Composio's managed implementation, you don't have to create your own developer app. For production, if you're building an end product, we recommend using your own credentials. The managed server helps you prototype fast and go from 0-1 faster. ## Step-by-step Guide ### 1. Prerequisites Before you begin, make sure you have: - Node.js and npm installed - A Composio account with API key - An OpenAI API key ### 1. Getting API Keys for OpenAI and Composio OpenAI API Key - Go to the [OpenAI dashboard](https://platform.openai.com/settings/organization/api-keys) 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](https://dashboard.composio.dev?utm_source=toolkits&utm_medium=framework_docs). - Navigate to your API settings and generate a new API key. - Store this key securely as you'll need it for authentication. ### 2. Install required dependencies 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 ```bash npm install @ai-sdk/openai @ai-sdk/mcp @composio/core ai dotenv ``` ### 3. Set up environment variables 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 ```bash OPENAI_API_KEY=your_openai_api_key_here COMPOSIO_API_KEY=your_composio_api_key_here COMPOSIO_USER_ID=your_user_id_here ``` ### 4. Import required modules and validate environment 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 ```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, }); ``` ### 5. Create Tool Router session and initialize MCP client What's happening: - We're creating a Tool Router session that gives your agent access to Bart 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 Bart-related tools through the MCP protocol ```typescript async function main() { // Create a tool router session for the user const session = await composio.create(composioUserID!, { toolkits: ["bart"], }); const mcpUrl = session.mcp.url; ``` ### 6. Connect to MCP server and retrieve 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 Bart tools that the agent can use ```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(); ``` ### 7. Initialize conversation and CLI interface 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 ```typescript let messages: ModelMessage[] = []; console.log("Chat started! Type 'exit' or 'quit' to end the conversation.\n"); console.log( "Ask any questions related to bart, like summarize my last 5 emails, send an email, etc... :)))\n", ); const rl = readline.createInterface({ input: process.stdin, output: process.stdout, prompt: "> ", }); rl.prompt(); ``` ### 8. Handle user input and stream responses with real-time tool feedback 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 Bart 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 ```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("\n๐Ÿ‘‹ Session ended."); process.exit(0); }); } main().catch((err) => { console.error("Fatal error:", err); process.exit(1); }); ``` ## Complete Code ```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: ["bart"], }); 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 bart, 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("\n๐Ÿ‘‹ Session ended."); process.exit(0); }); } main().catch((err) => { console.error("Fatal error:", err); process.exit(1); }); ``` ## Conclusion You've successfully built a Bart 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 Bart MCP Agent with another framework - [OpenAI Agents SDK](https://composio.dev/toolkits/bart/framework/open-ai-agents-sdk) - [Claude Agent SDK](https://composio.dev/toolkits/bart/framework/claude-agents-sdk) - [Claude Code](https://composio.dev/toolkits/bart/framework/claude-code) - [Claude Cowork](https://composio.dev/toolkits/bart/framework/claude-cowork) - [Codex](https://composio.dev/toolkits/bart/framework/codex) - [OpenClaw](https://composio.dev/toolkits/bart/framework/openclaw) - [Hermes](https://composio.dev/toolkits/bart/framework/hermes-agent) - [CLI](https://composio.dev/toolkits/bart/framework/cli) - [Google ADK](https://composio.dev/toolkits/bart/framework/google-adk) - [LangChain](https://composio.dev/toolkits/bart/framework/langchain) - [Mastra AI](https://composio.dev/toolkits/bart/framework/mastra-ai) - [LlamaIndex](https://composio.dev/toolkits/bart/framework/llama-index) - [CrewAI](https://composio.dev/toolkits/bart/framework/crew-ai) ## Related Toolkits - [Google Calendar](https://composio.dev/toolkits/googlecalendar) - Google Calendar is a time management service for scheduling meetings, events, and reminders. It streamlines personal and team organization with integrated notifications and sharing options. - [Apaleo](https://composio.dev/toolkits/apaleo) - Apaleo is a cloud-based property management platform for hospitality businesses. It centralizes reservations, billing, and daily operations for smoother hotel management. - [Appointo](https://composio.dev/toolkits/appointo) - Appointo is an appointment booking platform for Shopify stores. It lets businesses add online scheduling to their websites with zero coding. - [Bookingmood](https://composio.dev/toolkits/bookingmood) - Bookingmood is commission-free booking software for rental businesses. It lets you manage reservations and sync bookings directly on your website. - [Booqable](https://composio.dev/toolkits/booqable) - Booqable is a rental software platform for managing inventory, bookings, and reservations. It helps businesses streamline rentals and keep track of every item with ease. - [Cal](https://composio.dev/toolkits/cal) - Cal is a meeting scheduling platform that offers shareable booking links and real-time calendar syncing. It streamlines the process of finding mutual availability to make scheduling effortless. - [Calendarhero](https://composio.dev/toolkits/calendarhero) - Calendarhero is a powerful scheduling platform that streamlines your calendar management across multiple services. It helps you efficiently schedule, reschedule, and organize meetings without the back-and-forth. - [Calendly](https://composio.dev/toolkits/calendly) - Calendly is an appointment scheduling tool that automates meeting invitations, availability checks, and reminders. It helps individuals and teams avoid endless email back-and-forth when booking meetings. - [Etermin](https://composio.dev/toolkits/etermin) - eTermin is an online appointment scheduling platform for businesses to manage bookings. It streamlines client appointments, saving time and reducing scheduling conflicts. - [Evenium](https://composio.dev/toolkits/evenium) - Evenium is an all-in-one platform for managing professional events, from planning to analysis. It helps teams simplify event logistics, boost engagement, and track every detail in one place. - [Eventee](https://composio.dev/toolkits/eventee) - Eventee is a user-friendly event management platform for mobile and web. It boosts attendee engagement for in-person, virtual, and hybrid events. - [Eventzilla](https://composio.dev/toolkits/eventzilla) - Eventzilla is an event management platform for creating, promoting, and running events. It streamlines ticketing, registration, and attendee coordination for organizers. - [Humanitix](https://composio.dev/toolkits/humanitix) - Humanitix is a not-for-profit ticketing platform that donates 100% of profits to charity. It empowers event organizers to make social impact with every ticket sold. - [Lodgify](https://composio.dev/toolkits/lodgify) - Lodgify is an all-in-one vacation rental software for property managers and owners. It centralizes bookings, guest messaging, and channel synchronization in one dashboard. - [Planyo Online Booking](https://composio.dev/toolkits/planyo_online_booking) - Planyo Online Booking is a flexible reservation system for managing bookings by day, hour, or event. It streamlines scheduling for any business needing reservations. - [Scheduleonce](https://composio.dev/toolkits/scheduleonce) - Scheduleonce is a scheduling platform for capturing, qualifying, and engaging with inbound leads. It streamlines appointment booking and follow-ups for faster lead conversion. - [Supersaas](https://composio.dev/toolkits/supersaas) - Supersaas is a flexible appointment scheduling platform for businesses and individuals. It streamlines bookings, reminders, and calendar management in one place. - [Sympla](https://composio.dev/toolkits/sympla) - Sympla is a platform for managing in-person and online events, ticket sales, and registrations. It streamlines event setup, attendee tracking, and digital content delivery. - [Gmail](https://composio.dev/toolkits/gmail) - Gmail is Google's email service with powerful spam protection, search, and G Suite integration. It keeps your inbox organized and makes communication fast and reliable. - [Google Drive](https://composio.dev/toolkits/googledrive) - Google Drive is a cloud storage platform for uploading, sharing, and collaborating on files. It's perfect for keeping your documents accessible and organized across devices. ## Frequently Asked Questions ### What are the differences in Tool Router MCP and Bart MCP? With a standalone Bart MCP server, the agents and LLMs can only access a fixed set of Bart tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Bart 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 Bart tools. ### Can I manage the permissions and scopes for Bart while using Tool Router? Yes, absolutely. You can configure which Bart 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 Bart data and credentials are handled as safely as possible. --- [See all toolkits](https://composio.dev/toolkits) ยท [Composio docs](https://docs.composio.dev/llms.txt)