How to integrate Figma MCP with Vercel AI SDK

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

Introduction

This guide walks you through connecting Figma to Vercel AI SDK using the Composio tool router. By the end, you'll have a working Figma agent that can add a comment to this figma file, convert design tokens to tailwind css, delete a reaction from a comment, create a webhook for figma team events through natural language commands.

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

The Figma MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Figma account. It provides structured and secure access to your Figma workspace, so your agent can perform actions like commenting on designs, managing design tokens, linking developer resources, and automating collaboration workflows on your behalf.

  • Automated commenting and feedback loops: Have your agent add, reply to, or delete comments on Figma files and branches to streamline design reviews and team discussions.
  • Design token management and conversion: Let the agent extract, update, or convert design tokens in your files, including generating Tailwind CSS configurations for seamless dev handoff.
  • Developer resource integration: Automatically attach, update, or remove dev resources linked to Figma nodes, bridging the gap between design and development with contextual documentation or code references.
  • Webhook setup and automation: Enable your agent to create or delete webhooks for team events, making it easy to trigger notifications or workflows based on design activity.
  • Collaborative variable management: Empower the agent to batch-create, modify, or delete variables, collections, and modes across your design system, keeping everything consistent and up to date.

Supported Tools & Triggers

Tools
Add a comment to a filePosts a new comment to a figma file or branch, optionally replying to an existing root comment (replies cannot be nested); `region height` and `region width` in `client meta` must be positive if defining a comment region.
Add a reaction to a commentPosts a specified emoji reaction to an existing comment in a figma file or branch, requiring valid file key and comment id.
Create a webhookCreates a figma webhook for a `team id` to send post notifications for an `event type` to a publicly accessible https `endpoint`; an initial ping is sent unless `status` is `paused`.
Create dev resourcesCreates and attaches multiple uniquely-urled development resources to specified figma nodes, up to 10 per node.
Create, modify, or delete variablesManages variables, collections, modes, and their values in a figma file via batch create/update/delete operations; use temporary ids to link new related items in one request and ensure `variablemodevalues` match the target variable's `resolvedtype`.
Delete a commentDeletes a specific comment from a figma file or branch, provided the authenticated user is the original author of the comment.
Delete a reactionDeletes a specific emoji reaction from a comment in a figma file; the user must have originally created the reaction.
Delete a webhookPermanently deletes an existing webhook, identified by its unique `webhook id`; this operation is irreversible.
Delete dev resourceDeletes a development resource (used to link figma design elements to external developer information like code or tasks) from a specified figma file.
Design tokens to tailwindConvert design tokens to tailwind css configuration.
Detect backgroundDetect background layers for selected nodes.
Discover Figma Resources🔍 smart figma resource discovery - never guess ids again!
Download Figma ImagesDownload images from figma file nodes.
Extract design tokensExtract design tokens from figma files.
Extract Prototype InteractionsExtract prototype interactions and animations from figma files.
Get activity logsRetrieves activity log events from figma, allowing filtering by event types, time range, and pagination.
Get a webhookRetrieves detailed information about a specific webhook by its id, provided the webhook exists and is accessible to the user.
Get comments in a fileRetrieves all comments from an existing figma file, identified by a valid `file key`, returning details like content, author, position, and reactions, with an option for markdown formatted content.
Get componentGet component data with automatic simplification.
Get component setRetrieves detailed metadata for a specific published figma component set using its unique `key`.
Get current userRetrieves detailed information for the currently authenticated figma user.
Get dev resourcesRetrieves development resources (e.
Get file componentsRetrieves published components from a figma file, which must be a main file (not a branch) acting as a library.
Get file component setsRetrieves all published component sets from the specified figma main file (file key must not be for a branch).
Get file jsonGet figma file data with automatic simplification.
Get files in a projectFetches a list of files in a figma project, optionally including branch metadata.
Get file stylesRetrieves a list of published styles (like colors, text attributes, effects, and layout grids) from a specified main figma file (not a branch).
Get image fillsRetrieves temporary (14-day expiry) download urls for all image fills in a figma file; requires `imageref` from `paint` objects to map urls.
Get library analytics component action dataRetrieves component insertion and detachment analytics for a specified figma library, groupable by 'component' or 'team' and filterable by a date range (yyyy-mm-dd).
Get library analytics component usage dataRetrieves component usage analytics for a specified figma library file (identified by `file key`), with data groupable by 'component' or 'file'.
Get library analytics style action dataRetrieves style usage analytics (insertions, detachments) for a figma library, grouped by 'style' or 'team'; if providing a date range, ensure end date is not before start date.
Get library analytics style usage dataRetrieves style usage analytics for a figma library (specified by a valid `file key`), allowing data to be grouped by 'file' or 'style'.
Get library analytics variable action dataRetrieves weekly, paginated analytics data on variable insertions and detachments for a specified figma library (identified by `file key`), groupable by 'variable' or 'team', and filterable by an optional date range.
Get library analytics variable usage dataRetrieves paginated analytics data on variable usage from a specified figma library, grouped by 'file' or 'variable', for libraries with enabled analytics.
Get local variablesRetrieves all local/remote variables for a figma file/branch; crucial for obtaining mode-specific values which `/v1/files/{file key}/variables/published` omits.
Get paymentsRetrieves a user's payment information for a figma plugin, widget, or community file; the authenticated identity must own the resource.
Get projects in a teamRetrieves projects within a specified figma team that are visible to the authenticated user.
Get published variablesRetrieves variables published from a specified figma file; this api is available only to full members of enterprise organizations.
Get reactions for a commentRetrieves reactions for a specific comment in a figma file.
Get styleRetrieves detailed metadata for a specific style in figma using its unique style key.
Get team componentsRetrieves components published in a specific figma team's library; the team must have published components, otherwise an empty list is returned.
Get team component setsRetrieves a paginated list of published component sets (collections of reusable ui elements) from a specified figma team's library.
Get team stylesRetrieves a paginated list of published styles, such as colors or text attributes, from a specified figma team's library.
Get team webhooksRetrieves all webhooks registered for a specified figma team.
Get versions of a fileRetrieves the version history for a figma file or branch, as specified by its `file key`.
Get webhook requestsRetrieves a history of webhook requests for a specific figma webhook subscription; data is available for requests sent within the last seven days.
Render images of file nodesRenders specified nodes from a figma file as images (jpg, pdf, png, svg), returning a map of node ids to image urls (or `null` for failed nodes); images expire after 30 days and are capped at 32 megapixels (larger requests are scaled down).
Update a webhookUpdates an existing figma webhook, identified by `webhook id`, allowing modification of its event type, endpoint, passcode, status, or description.
Update dev resourcesUpdates the name and/or url of one or more existing figma dev resources, each identified by its unique `id`.

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: ["figma"],
  });

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

  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 figma, 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 Figma 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 Figma MCP Agent with another framework

OpenAI Agents SDK

OpenAI Agents SDK

Use Figma MCP with OpenAI Agents SDK

Claude Agent SDK

Claude Agent SDK

Use Figma MCP with Claude Agent SDK

Google ADK

Google ADK

Use Figma MCP with Google ADK

LangChain

LangChain

Use Figma MCP with LangChain

Mastra AI

Mastra AI

Use Figma MCP with Mastra AI

LlamaIndex

LlamaIndex

Use Figma MCP with LlamaIndex

CrewAI

CrewAI

Use Figma MCP with CrewAI

FAQ

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

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

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

Yes, absolutely. You can configure which Figma 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 Figma 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