How to integrate Zulip MCP with Vercel AI SDK v6

GET STARTED FOR FREE
Framework Integration Gradient
Zulip Logo
Vercel AI SDK Logo
divider

Introduction

This guide walks you through connecting Zulip to Vercel AI SDK v6 using the Composio tool router. By the end, you'll have a working Zulip agent that can post daily standup summary to #team stream, list unread messages from #support stream, reply to the latest thread in #engineering through natural language commands.

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

The Zulip MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Zulip account. It provides structured and secure access so your agent can perform Zulip operations on your behalf.

Supported Tools & Triggers

Tools
Add Alert WordsTool to add alert words to the current user's configuration.
Add APNs Device TokenTool to add an APNs (Apple Push Notification service) device token to Zulip.
Add Default StreamTool to add a channel as a default stream in Zulip.
Add LinkifierTool to add a linkifier to a Zulip realm.
Add Navigation ViewTool to add a navigation view in Zulip.
Add ReactionTool to add an emoji reaction to a Zulip message.
Archive ChannelTool to archive a channel in Zulip.
Check Messages Match NarrowTool to check if specific messages match a narrow filter.
Create BigBlueButton video callTool to create a BigBlueButton video call in Zulip.
Create ChannelTool to create a new channel in Zulip.
Create Channel FolderTool to create a new channel folder in Zulip.
Create custom profile fieldTool to create a custom profile field in Zulip.
Create DraftsTool to create one or more draft messages in Zulip.
Create reusable invitation linkTool to create a reusable invitation link for the Zulip organization.
Create Message ReminderTool to create a message reminder in Zulip.
Create Saved SnippetTool to create a saved snippet in Zulip.
Create Scheduled MessageTool to create a scheduled message in Zulip.
Delete DraftTool to delete a draft in Zulip.
Delete MessageTool to delete a message in Zulip.
Delete Event QueueTool to delete an event queue previously registered via POST /api/v1/register.
Delete ReminderTool to delete a reminder from Zulip.
Delete Saved SnippetTool to delete a saved snippet by its ID.
Delete Scheduled MessageTool to delete a scheduled message in Zulip.
Delete TopicTool to delete a topic from a Zulip channel/stream.
Edit DraftTool to edit an existing draft message in Zulip.
Edit Navigation ViewTool to update a navigation view in Zulip.
Edit Saved SnippetTool to edit a saved snippet by updating its title and/or content.
Export RealmTool to create a Zulip realm data export.
Get alert wordsTool to retrieve all configured alert words for the authenticated user.
Get Another User PresenceTool to get the presence status for a specific user in Zulip.
Get AttachmentsTool to retrieve metadata on files uploaded by the requesting user.
Get Channel FoldersTool to retrieve all channel folders in the organization, sorted by display order.
Get custom emojiTool to retrieve all custom emoji configured in the Zulip realm.
Get custom profile fieldsTool to get all custom profile fields configured for the Zulip realm.
Get draftsTool to fetch all drafts for the authenticated user.
Get EventsTool to get events from a Zulip event queue.
Get file temporary URLTool to get a public temporary URL for an uploaded file in Zulip.
Get realm iconTool to retrieve the URL of the Zulip realm's icon.
Get all invitationsTool to retrieve all unexpired invitations for the Zulip organization.
Get user group membership statusTool to check if a user is a member of a specific user group.
Get linkifiersTool to retrieve all configured linkifiers in the Zulip realm.
Get messageTool to fetch a single message by ID from Zulip.
Get Message HistoryTool to fetch the complete edit history of a Zulip message.
Get MessagesTool to fetch messages from Zulip.
Get Navigation ViewsTool to fetch all navigation views for the authenticated user.
Get own userTool to retrieve the profile information for the authenticated user.
Get Read ReceiptsTool to retrieve read receipts for a message in Zulip.
Get realm export consentsTool to retrieve data export consent state for users in a realm.
Get all data exportsTool to retrieve all data exports for the Zulip organization.
Get Realm PresenceTool to get the presence status for all users in the Zulip realm.
Get remindersTool to fetch all undelivered reminders for the authenticated user.
Get Saved SnippetsTool to fetch all saved snippets for the authenticated user.
Get Scheduled MessagesTool to fetch all scheduled messages for the authenticated user.
Get server settingsTool to retrieve server configuration and available authentication methods.
Get Channel by IDTool to get detailed information about a Zulip channel by its ID.
Get channel email addressTool to retrieve the email address for a Zulip channel.
Get Channel IDTool to get a channel's unique ID by its name.
Get All ChannelsTool to retrieve all channels (streams) in Zulip that match specified filters.
Get stream topicsTool to retrieve all topics in a Zulip channel.
Get Channel SubscribersTool to get all subscribers of a Zulip channel.
Get SubscriptionsTool to retrieve all channels (streams) the authenticated user is subscribed to.
Get subscription statusTool to check whether a specific user is subscribed to a particular channel (stream).
Get user by emailTool to fetch a user by email address from Zulip.
Get user channelsTool to retrieve the list of channel IDs that a specific user is subscribed to.
Get user group membersTool to retrieve all members of a specific user group.
Get user groupsTool to retrieve all user groups in the Zulip organization.
Get user group subgroupsTool to retrieve subgroups of a user group.
Get user statusTool to retrieve the status currently set by a user in the Zulip organization.
Mark All Messages as ReadTool to mark all messages as read for the authenticated user.
Mark Stream as ReadTool to mark all messages in a Zulip channel as read.
Mark Topic as ReadTool to mark all messages in a topic as read.
Mute TopicTool to mute or unmute a topic in a Zulip channel.
Reactivate UserTool to reactivate a deactivated user in Zulip.
Register Push DeviceTool to register an E2EE (end-to-end encrypted) push device for mobile notifications.
Register Event QueueTool to register an event queue for receiving Zulip events.
Remove Alert WordsTool to remove alert words from the current user's configuration.
Remove AttachmentTool to delete an attachment in Zulip.
Remove Code PlaygroundTool to remove a code playground from a Zulip realm.
Remove Default StreamTool to remove a channel from default streams in Zulip.
Remove FCM TokenTool to remove an FCM registration token for Android push notifications.
Remove LinkifierTool to remove a linkifier from a Zulip realm.
Remove Navigation ViewTool to remove a navigation view in Zulip.
Remove ReactionTool to remove an emoji reaction from a Zulip message.
Render MessageTool to render a message to HTML in Zulip.
Reorder Channel FoldersTool to reorder channel folders in Zulip.
Reorder Custom Profile FieldsTool to reorder custom profile fields in Zulip.
Reorder LinkifiersTool to reorder linkifiers in Zulip.
Revoke Email InvitationTool to revoke an email invitation in Zulip.
Revoke reusable invitation linkTool to revoke a reusable invitation link for the Zulip organization.
Send MessageTool to send a message in Zulip.
Send Test NotificationTool to send a test notification to mobile device(s).
Set Typing StatusTool to set typing status in Zulip.
Set Typing Status for Message EditTool to set typing status when editing a message in Zulip.
Subscribe to ChannelTool to subscribe users to one or more Zulip channels.
Test Welcome Bot Custom MessageTool to test welcome bot custom message in Zulip.
Unsubscribe From ChannelTool to unsubscribe from one or more channels (streams) in Zulip.
Update Channel FolderTool to update a channel folder in Zulip.
Update LinkifierTool to update a linkifier in a Zulip realm.
Update MessageTool to edit a message in Zulip.
Update Message FlagsTool to update personal message flags in Zulip.
Update Message Flags for NarrowTool to update personal message flags (e.
Update Realm User Settings DefaultsTool to update realm-level defaults of user settings in Zulip.
Update Scheduled MessageTool to edit a scheduled message in Zulip.
Update SettingsTool to update user settings in Zulip.
Update StatusTool to update your status in Zulip.
Update status for userTool to update the status for a specific user in the Zulip organization.
Update ChannelTool to update a channel in Zulip.
Update SubscriptionsTool to update channel subscriptions by subscribing to new channels or unsubscribing from existing ones.
Update Subscription SettingsTool to update the user's personal settings for channels they are subscribed to.
Update user by emailTool to update a user by email address in Zulip.
Update User TopicTool to update personal preferences for a topic in a Zulip channel.

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 { 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: ["zulip"],
  });

  const mcpUrl = session.mcp.url;
What's happening:
  • We're creating a Tool Router session that gives your agent access to Zulip 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 Zulip-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 Zulip 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 zulip, 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("\nđź‘‹ 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 Zulip 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 Zulip 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: ["zulip"],
  });

  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 zulip, 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("\nđź‘‹ Session ended.");
    process.exit(0);
  });
}

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

Conclusion

You've successfully built a Zulip 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 Zulip MCP Agent with another framework

OpenAI Agents SDK

OpenAI Agents SDK

Use Zulip MCP with OpenAI Agents SDK

Claude Agent SDK

Claude Agent SDK

Use Zulip MCP with Claude Agent SDK

Google ADK

Google ADK

Use Zulip MCP with Google ADK

LangChain

LangChain

Use Zulip MCP with LangChain

Mastra AI

Mastra AI

Use Zulip MCP with Mastra AI

LlamaIndex

LlamaIndex

Use Zulip MCP with LlamaIndex

CrewAI

CrewAI

Use Zulip MCP with CrewAI

Claude Code

Claude Code

Use Zulip MCP with Claude Code

Codex

Codex

Use Zulip MCP with Codex

FAQ

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

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

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

Yes, absolutely. You can configure which Zulip 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 Zulip 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.

Book a call↗Get started for free