How to integrate Coda MCP with Vercel AI SDK

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

Introduction

This guide walks you through connecting Coda to Vercel AI SDK using the Composio tool router. By the end, you'll have a working Coda agent that can duplicate my project tracker document, add a new permission for this doc, export the content of the roadmap page, create a new page in q2 planning doc through natural language commands.

This guide will help you understand how to give your Vercel AI SDK agent real control over a Coda account through Composio's Coda 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 Coda integration
  • Using Composio's Tool Router to dynamically load and access Coda tools
  • Creating an MCP client connection using HTTP transport
  • Building an interactive CLI chat interface with conversation history management
  • Handling tool calls and results within the Vercel AI SDK framework

What is Vercel AI SDK?

The Vercel AI SDK is a TypeScript library for building AI-powered applications. It provides tools for creating agents that can use external services and maintain conversation state.

Key features include:

  • streamText: Core function for streaming responses with real-time tool support
  • MCP Client: Built-in support for Model Context Protocol
  • Step Counting: Control multi-step tool execution
  • OpenAI Provider: Native integration with OpenAI models

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

The Coda MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Coda account. It provides structured and secure access to your workspaces and docs, so your agent can automate Coda document creation, manage permissions, export content, and streamline your team’s productivity tools—all on your behalf.

  • Automated document and page creation: Instruct your agent to create new Coda documents or pages, duplicate existing docs, and organize content structure with just a prompt.
  • Flexible permission and sharing management: Have your agent add or modify user, workspace, or global permissions, so you’re always in control of who can view or edit your docs.
  • Seamless content export and status tracking: Let your agent initiate exports of Coda pages and check the progress, making it easy to share or archive important information.
  • Custom domain and publishing management: Direct your agent to add custom domains to published docs or manage categories and makers for Coda packs, keeping your workspace organized and discoverable.
  • Pack release and workspace enhancements: Ask your agent to create new pack releases or manage integrations, helping teams extend Coda’s power without repetitive manual steps.

Supported Tools & Triggers

Tools
Triggers
Add a category for packAdd a publishing category for a given pack.
Add a maker for packSet a maker for a given pack.
Add a permission for packCreate or modify user, workspace, or global permissions for a given pack.
Add custom domainAdd a custom domain to a published doc.
Add permissionAdds a new permission to the doc.
Begin content exportInitiate an export of content for the given page.
Content export statusCheck the status of a page content export
Copy DocumentCreates a copy of an existing coda document.
Create a new pack releaseCreates a new pack release based on an existing pack version.
Create a pageCreate a new page in a doc.
Create docCreates a new coda doc, optionally copying an existing doc.
Create packCreates a new pack, essentially registering a new pack id.
Delete a category for packDelete a publishing category for a given pack.
Delete a maker for packDelete a maker for a given pack, who will not be displayed in the corresponding packs page.
Delete a pageDeletes the specified page.
Delete a permission for packDelete user, workspace, or global permissions for a given pack.
Delete docDeletes a doc.
Delete multiple rowsDeletes the specified rows from the table or view.
Delete packDelete a given pack.
Delete permissionDeletes an existing permission.
Delete rowDeletes the specified row from the table or view.
Deletes a custom domainDeletes a custom domain from a published doc.
Fetch grouped logs by pack org root ingestion idRetrieve the grouped logs of a pack for debugging purpose.
Fetch ingestion executions for packRetrieve the ingestion execution ids of a root ingestion for debugging purpose.
Get acl settingsReturns settings associated with acls for this coda doc.
Get a columnReturns details about a column in a table.
Get a controlReturns info on a control.
Get a formulaReturns info on a formula.
Get analytics last updated dayReturns days based on pacific standard time when analytics were last updated.
Get a pageReturns details about a page.
Get a rowReturns details about a row in a table.
Get a single packReturns a single pack.
Get a tableReturns details about a specific table or view.
Get detailed listing information for a packGet detailed listing information for a pack.
Get doc analytics summaryReturns summarized analytics data for available docs.
Get doc categoriesGets all available doc categories.
Get info about a docReturns metadata for the specified doc.
Get mutation statusGet the status for an asynchronous mutation to know whether or not it has been completed.
Get pack analytics summaryReturns summarized analytics data for packs the user can edit.
Gets custom doc domains providersGets the provider (ie.
Get sharing metadataReturns metadata associated with sharing for this coda doc.
Gets the json schema for pack configurationReturns a json schema applicable for customizing the pack using pack configurations.
Get the difference between two pack versionsGets information about the difference between the specified previous version and next version of a pack.
Get the next valid version for a packGet the next valid version based on the proposed metadata.
Get the source code for a pack versionGet temporary links used to download the source code for the given packid and version
Get user infoReturns basic info about the current user.
List available docsReturns a list of coda docs accessible by the user, and which they have opened at least once.
List categories for packList publishing categories for a given pack.
List columnsReturns a list of columns in a table.
List controlsReturns a list of controls in a coda doc.
List custom doc domainsList all custom domains for a published doc.
List doc analyticsReturns analytics data for available docs per day.
List featured docs for a packReturns a list of featured doc ids for a pack.
List formulasReturns a list of named formulas in a coda doc.
List makers for packList makers for a given pack.
List pack analyticsReturns analytics data for packs the user can edit.
List pack formula analyticsReturns analytics data for pack formulas.
List packsGet the list of accessible packs.
List page analyticsReturns analytics data for a given doc within the day.
List pagesReturns a list of pages in a coda doc.
List permissionsReturns a list of permissions for this coda doc.
List permissions for a packGet user, workspace, and/or global permissions for a given pack.
List table rowsRetrieves rows from a specific table within a coda document.
List tablesReturns a list of tables in a coda doc.
List the pack listings accessible to a userGet listings of public packs and packs created by you.
List the releases for a packGet the list of releases of a pack.
List the versions for a packGet the list of versions of a pack.
List workspace rolesReturns a list of the counts of users over time by role for the workspace.
List workspace usersReturns a list of members in the given workspace.
Pack asset upload completeNote the completion of the upload of a pack asset.
Pack source code upload completeNote the completion of the upload of a pack source code.
Pack version upload completeNote the completion of the upload of a pack version bundle in order to create that pack version.
Patch the system connection credentials of the packPatch the system connection credentials of the pack.
Publish docUpdate publish settings for a doc.
Push a buttonPushes a button on a row in a table.
Register pack versionRegisters a new pack version.
Resolve browser linkGiven a browser link to a coda object, attempts to find it and return metadata that can be used to get more info on it.
Retrieve the grouped logs of a packRetrieve the grouped logs of a pack for debugging purpose.
Retrieve the information for a specific logRetrieve the ingestion execution ids of a root ingestion for debugging purpose.
Retrieve the logs of a ingestionRetrieve the logs of a ingestion for debugging purpose.
Retrieve the logs of a packRetrieve the logs of a pack for debugging purpose.
Retrieve the oauth configuration of the packRetrieve the oauth configuration of the pack for display purpose.
Retrieve the system connection metadata of the packRetrieve the system connection metadata of the pack.
Search Coda DocumentsThis tool allows users to search for coda documents based on a query term.
Search principalsSearches for user and group principals matching the query that this doc can be shared with.
Search Table RowsAction to search for rows in a coda table based on specific criteria.
Set the oauth configurations of the packSet the oauth configurations of the pack, including client id and secret.
Set the system connection credentials of the packSet the system connection credentials of the pack.
Trigger automationTriggers webhook-invoked automation
Unpublish docUnpublishes a doc.
Update acl settingsUpdate settings associated with acls for this coda doc.
Update an existing pack releaseUpdate details of a pack release.
Update a pageUpdate properties for a page.
Update docUpdates metadata for a doc.
Update featured docs for a packCreate or replace the featured docs for a pack.
Update packUpdate an existing pack for non-versioned fields.
Updates a custom domainUpdates properties of a document's custom domain.
Updates user roleUpdates the workspace user role of a user that matches the parameters.
Upload a pack assetRequest a signed s3 url to upload your pack asset.
Upload pack source codeRequest a signed s3 url to upload your pack source code.
Insert/Update Rows in Coda TableThis tool allows you to insert new rows into a coda table or update existing ones based on specified key columns.

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

What is Tool Router?

Composio's Tool Router helps agents find the right tools for a task at runtime. You can plug in multiple toolkits (like Gmail, HubSpot, and GitHub), and the agent will identify the relevant app and action to complete multi-step workflows. This can reduce token usage and improve the reliability of tool calls. Read more here: Getting started with Tool Router

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

How the Tool Router works

The Tool Router follows a three-phase workflow:

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

Step-by-step Guide

Prerequisites

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

Getting API Keys for OpenAI and Composio

OpenAI API Key
  • Go to the OpenAI dashboard and create an API key. You'll need credits to use the models, or you can connect to another model provider.
  • Keep the API key safe.
Composio API Key
  • Log in to the Composio dashboard.
  • Navigate to your API settings and generate a new API key.
  • Store this key securely as you'll need it for authentication.

Install required dependencies

bash
npm install @ai-sdk/openai @ai-sdk/mcp @composio/core ai dotenv

First, install the necessary packages for your project.

What you're installing:

  • @ai-sdk/openai: Vercel AI SDK's OpenAI provider
  • @ai-sdk/mcp: MCP client for Vercel AI SDK
  • @composio/core: Composio SDK for tool integration
  • ai: Core Vercel AI SDK
  • dotenv: Environment variable management

Set up environment variables

bash
OPENAI_API_KEY=your_openai_api_key_here
COMPOSIO_API_KEY=your_composio_api_key_here
COMPOSIO_USER_ID=your_user_id_here

Create a .env file in your project root.

What's needed:

  • OPENAI_API_KEY: Your OpenAI API key for GPT model access
  • COMPOSIO_API_KEY: Your Composio API key for tool access
  • COMPOSIO_USER_ID: A unique identifier for the user session

Import required modules and validate environment

typescript
import "dotenv/config";
import { openai } from "@ai-sdk/openai";
import { Composio } from "@composio/core";
import * as readline from "readline";
import { streamText, type ModelMessage, stepCountIs } from "ai";
import { experimental_createMCPClient as createMCPClient } from "@ai-sdk/mcp";

const composioAPIKey = process.env.COMPOSIO_API_KEY;
const composioUserID = process.env.COMPOSIO_USER_ID;

if (!process.env.OPENAI_API_KEY) throw new Error("OPENAI_API_KEY is not set");
if (!composioAPIKey) throw new Error("COMPOSIO_API_KEY is not set");
if (!composioUserID) throw new Error("COMPOSIO_USER_ID is not set");

const composio = new Composio({
  apiKey: composioAPIKey,
});
What's happening:
  • We're importing all necessary libraries including Vercel AI SDK's OpenAI provider and Composio
  • The dotenv/config import automatically loads environment variables
  • The MCP client import enables connection to Composio's tool server

Create Tool Router session and initialize MCP client

typescript
async function main() {
  // Create a tool router session for the user
  const { session } = await composio.create(composioUserID!, {
    toolkits: ["coda"],
  });

  const mcpUrl = session.mcp.url;
What's happening:
  • We're creating a Tool Router session that gives your agent access to Coda 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 Coda-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 Coda 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 coda, 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 Coda 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 Coda and Vercel AI SDK:

typescript
import "dotenv/config";
import { openai } from "@ai-sdk/openai";
import { Composio } from "@composio/core";
import * as readline from "readline";
import { streamText, type ModelMessage, stepCountIs } from "ai";
import { experimental_createMCPClient as createMCPClient } from "@ai-sdk/mcp";

const composioAPIKey = process.env.COMPOSIO_API_KEY;
const composioUserID = process.env.COMPOSIO_USER_ID;

if (!process.env.OPENAI_API_KEY) throw new Error("OPENAI_API_KEY is not set");
if (!composioAPIKey) throw new Error("COMPOSIO_API_KEY is not set");
if (!composioUserID) throw new Error("COMPOSIO_USER_ID is not set");

const composio = new Composio({
  apiKey: composioAPIKey,
});

async function main() {
  // Create a tool router session for the user
  const { session } = await composio.create(composioUserID!, {
    toolkits: ["coda"],
  });

  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 coda, 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 Coda 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 Coda MCP Agent with another framework

OpenAI Agents SDK

OpenAI Agents SDK

Use Coda MCP with OpenAI Agents SDK

Claude Agent SDK

Claude Agent SDK

Use Coda MCP with Claude Agent SDK

Google ADK

Google ADK

Use Coda MCP with Google ADK

LangChain

LangChain

Use Coda MCP with LangChain

Mastra AI

Mastra AI

Use Coda MCP with Mastra AI

LlamaIndex

LlamaIndex

Use Coda MCP with LlamaIndex

CrewAI

CrewAI

Use Coda MCP with CrewAI

FAQ

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

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

Can I use Tool Router MCP with Vercel AI SDK?

Yes, you can. Vercel AI SDK fully supports MCP integration. You get structured tool calling, message history handling, and model orchestration while Tool Router takes care of discovering and serving the right Coda tools.

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

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

Used by agents from

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

Never worry about agent reliability

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

Book a call↗Get started for free