How to integrate Daytona MCP with Claude Agent SDK

This guide walks you through connecting Daytona to the Claude Agent SDK using the Composio tool router. By the end, you'll have a working Daytona agent that can list daytona workspaces and current status, create cloud workspace for node project, stop idle daytona development workspaces through natural language commands. This guide will help you understand how to give your Claude Agent SDK agent real control over a Daytona account through Composio's Daytona MCP server. Before we dive in, let's take a quick look at the key ideas and tools involved.

Daytona logoDaytona
Api KeyOauth2

Daytona is a development environment management platform for cloud-based workspaces. Use it to provision, manage, and automate dev environments through APIs.

114 Tools

Introduction

This guide walks you through connecting Daytona to the Claude Agent SDK using the Composio tool router. By the end, you'll have a working Daytona agent that can list daytona workspaces and current status, create cloud workspace for node project, stop idle daytona development workspaces through natural language commands.

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

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

Also integrate Daytona with

TL;DR

Here's what you'll learn:
  • Get and set up your Claude/Anthropic and Composio API keys
  • Install the necessary dependencies
  • Initialize Composio and create a Tool Router session for Daytona
  • Configure an AI agent that can use Daytona as a tool
  • Run a live chat session where you can ask the agent to perform Daytona operations

What is Claude Agent SDK?

The Claude Agent SDK is Anthropic's official framework for building AI agents powered by Claude. It provides a streamlined interface for creating agents with MCP tool support and conversation management.

Key features include:

  • Native MCP Support: Built-in support for Model Context Protocol servers
  • Permission Modes: Control tool execution permissions
  • Streaming Responses: Real-time response streaming for interactive applications
  • Context Manager: Clean async context management for sessions

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

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

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

What is Composio SDK?

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

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

How the Composio SDK works

The Composio SDK follows a three-phase workflow:

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

Step-by-step Guide

Step by step09 STEPS
1

Prerequisites

Before starting, make sure you have:
  • Composio API Key and Claude/Anthropic API Key
  • Primary know-how of Claude Agents SDK
  • A Daytona account
  • Some knowledge of Python
2

Getting API Keys for Claude/Anthropic and Composio

Claude/Anthropic API Key
  • Go to the Anthropic Console and create an API key. You'll need credits to use the models.
  • 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.
3

Install dependencies

npm install @anthropic-ai/claude-agent-sdk @composio/core dotenv

Install the Composio SDK and the Claude Agents SDK.

What's happening:

  • @composio/core provides Composio integration for Anthropic
  • @anthropic-ai/claude-agent-sdk is the core agent framework
  • dotenv/config loads environment variables
4

Set up environment variables

bash
COMPOSIO_API_KEY=your_composio_api_key_here
USER_ID=your_user_id_here
ANTHROPIC_API_KEY=your_anthropic_api_key_here

Create a .env file in your project root.

What's happening:

  • COMPOSIO_API_KEY authenticates with Composio
  • USER_ID identifies the user for session management
  • ANTHROPIC_API_KEY authenticates with Anthropic/Claude
5

Import dependencies

import 'dotenv/config';
import readline from 'node:readline';
import { Composio } from '@composio/core';
import { query, type Options } from "@anthropic-ai/claude-agent-sdk";

dotenv.config();
What's happening:
  • We're importing all necessary libraries including the Claude Agent SDK and Composio
  • The dotenv.config() function loads environment variables from your .env file
  • This setup prepares the foundation for connecting Claude with Daytona functionality
6

Create a Composio instance and Tool Router session

async function chat() {
  const { COMPOSIO_API_KEY, USER_ID } = process.env;
  if (!COMPOSIO_API_KEY || !USER_ID) {
    throw new Error('COMPOSIO_API_KEY and USER_ID required in .env');
  }

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

  // Create Tool Router session for Daytona
  const session = await composio.create(USER_ID, {
    toolkits: ['daytona'],
  });
  const mcpUrl = session?.mcp.url;
What's happening:
  • The function checks for the required COMPOSIO_API_KEY environment variable
  • We're creating a Composio instance using our API key
  • The create method creates a Tool Router session for Daytona
  • The returned url is the MCP server URL that your agent will use
7

Configure Claude Agent with MCP

const options: Options = {
  permissionMode: 'bypassPermissions',
  mcpServers: {
    composio: {
      type: 'http',
      url: mcpUrl,
      headers: { 'x-api-key': COMPOSIO_API_KEY }
    }
  },
  systemPrompt: 'You are a helpful assistant with access to Daytona tools via Composio.',
  maxTurns: 10,
};
What's happening:
  • We're configuring the Claude Agent options with the MCP server URL
  • permissionMode: 'bypassPermissions' allows the agent to execute operations without asking for permission each time
  • The system prompt instructs the agent that it has access to Daytona
  • maxTurns: 10 limits the conversation length to prevent excessive API usage
8

Create client and start chat loop

const rl = readline.createInterface({
    input: process.stdin,
    output: process.stdout,
    prompt: 'You: '
  });

  console.log('\nChat started. Type "exit" to quit.\n');

  let isProcessing = false;

  async function ask(prompt: string) {
    isProcessing = true;
    rl.pause();

    process.stdout.write('Claude is thinking...');
    const stream = query({ prompt, options });

    let firstChunk = true;
    for await (const msg of stream) {
      const content = (msg as any).message?.content || (msg as any).content;
      if (Array.isArray(content)) {
        for (const block of content) {
          if (block.type === 'text' && block.text) {
            if (firstChunk) {
              process.stdout.write('\r\x1b[K');
              process.stdout.write('Claude: ');
              firstChunk = false;
            }
            process.stdout.write(block.text);
          }
        }
      }
    }
    process.stdout.write('\n\n');

    isProcessing = false;
    rl.resume();
    rl.prompt();
  }

  rl.on('line', async (line) => {
    if (isProcessing) return;

    const input = line.trim();
    if (input === 'exit') {
      rl.close();
      process.exit(0);
    }
    if (input) await ask(input);
    else rl.prompt();
  });

  await ask('What can you help me with?');
}
What's happening:
  • The readline interface is created to handle user input and output
  • The query function is used to send the user's input to the agent
  • The chat loop continues until the user types 'exit' or 'quit'
9

Run the application

try {
  await chat();
} catch (error) {
  console.error(error);
  process.exit(1);
}
What's happening:
  • The chat function is the entry point for the application
  • The try-catch block is used to handle any errors that occur

Complete Code

Here's the complete code to get you started with Daytona and Claude Agent SDK:

import 'dotenv/config';
import readline from 'node:readline';
import { Composio } from '@composio/core';
import { query, type Options } from "@anthropic-ai/claude-agent-sdk";

async function chat() {
  const { COMPOSIO_API_KEY, USER_ID } = process.env;
  if (!COMPOSIO_API_KEY || !USER_ID) {
    throw new Error('COMPOSIO_API_KEY and USER_ID required in .env');
  }

  const composio = new Composio({ apiKey: COMPOSIO_API_KEY });
  const session = await composio.create(USER_ID, {
    toolkits: ['daytona']
  });
  const mcp_url = session?.mcp.url;

  const options: Options = {
    permissionMode: 'bypassPermissions',
    mcpServers: {
      composio: {
        type: 'http',
        url: mcp_url,
        headers: { 'x-api-key': COMPOSIO_API_KEY }
      }
    },
    systemPrompt: 'You are a helpful assistant with access to Daytona tools via Composio.',
    maxTurns: 10,
  };

  const rl = readline.createInterface({
    input: process.stdin,
    output: process.stdout,
    prompt: 'You: '
  });

  console.log('\nChat started. Type "exit" to quit.\n');

  let isProcessing = false;

  async function ask(prompt: string) {
    isProcessing = true;
    rl.pause();

    process.stdout.write('Claude is thinking...');
    const stream = query({ prompt, options });

    let firstChunk = true;
    for await (const msg of stream) {
      const content = (msg as any).message?.content || (msg as any).content;
      if (Array.isArray(content)) {
        for (const block of content) {
          if (block.type === 'text' && block.text) {
            if (firstChunk) {
              process.stdout.write('\r\x1b[K');
              process.stdout.write('Claude: ');
              firstChunk = false;
            }
            process.stdout.write(block.text);
          }
        }
      }
    }
    process.stdout.write('\n\n');

    isProcessing = false;
    rl.resume();
    rl.prompt();
  }

  rl.on('line', async (line) => {
    if (isProcessing) return;

    const input = line.trim();
    if (input === 'exit') {
      rl.close();
      process.exit(0);
    }
    if (input) await ask(input);
    else rl.prompt();
  });

  await ask('What can you help me with?');
}

try {
  await chat();
} catch (error) {
  console.error(error);
  process.exit(1);
}

Conclusion

You've successfully built a Claude Agent SDK agent that can interact with Daytona through Composio's Tool Router.

Key features:

  • Native MCP support through Claude's agent framework
  • Streaming responses for real-time interaction
  • Permission bypass for smooth automated workflows
You can extend this by adding more toolkits, implementing custom business logic, or building a web interface around the agent.
TOOLS

Supported Tools

Every Daytona action and event your agent gets out of the box.

Archive Sandbox

Archives a sandbox in the Daytona platform.

Cancel Organization Invitation

Cancels a pending organization invitation.

Checkout Git Branch

[DEPRECATED] Checkout a branch or commit in a git repository within a sandbox.

Clone Repository

Clones a Git repository into a sandbox toolbox.

Create API Key

Creates a new API key for the organization.

Create Sandbox Backup

Initiates a backup operation for an existing sandbox environment.

Create Folder in Sandbox

Creates a new folder inside a sandbox's filesystem.

Create Organization

Creates a new organization in Daytona.

Create Organization Invitation

Creates an invitation to join an organization with specified roles and permissions.

Create Organization Role

Creates a new organization role with specified permissions.

Create Process Session

Creates a new session in the sandbox for process management.

Create PTY Session

Creates a new PTY (pseudo-terminal) session in a sandbox.

Create Docker Registry

Creates a new Docker registry for the organization.

Create Sandbox

Creates a new sandbox environment for development or testing purposes.

Create Snapshot

Creates a new snapshot resource that can be used as a base image for sandboxes.

Create SSH Access

Creates SSH access credentials for a sandbox environment.

Create Volume

Creates a new volume for storing data that can be attached to sandboxes.

Deactivate Snapshot

Deactivates a snapshot in the Daytona platform.

Delete API Key

Deletes an API key by its name.

Delete API Key for User

Deletes an API key for a specific user.

Delete Sandbox File

Deletes a file inside a sandbox by its path.

Delete Git Branch

Deletes a branch from a git repository inside a sandbox by its name.

Delete Organization

Deletes an organization.

Delete Organization OTEL Config

Deletes the OpenTelemetry (OTEL) configuration for an organization.

Delete Organization Role

Deletes an organization role.

Delete PTY Session

Deletes a PTY session and terminates the associated process.

Delete Recording

Deletes a recording from the Daytona platform by its ID.

Delete Docker Registry

Deletes a Docker registry from the Daytona platform.

Delete Sandbox

Deletes a sandbox from the Daytona platform by its ID or name.

Delete Session

Deletes a session from the Daytona platform by its ID.

Delete Snapshot

Deletes a snapshot from the Daytona platform by its ID.

Delete Volume

Deletes a volume from the Daytona platform by its ID.

Download File (Deprecated)

[DEPRECATED] Downloads a file from a sandbox's toolbox filesystem.

Download Multiple Files

Downloads multiple files from a sandbox as a multipart/form-data bundle.

Enroll in SMS MFA

Enrolls the authenticated user in SMS-based multi-factor authentication.

Execute Command

Executes a shell command within a running sandbox via the toolbox proxy.

Expire Signed Port Preview URL

Expires a signed preview URL for a sandbox port.

Find In Files (Deprecated)

[DEPRECATED] Searches for text or patterns in files within a sandbox.

Focus Accessibility Node

Focuses an accessibility node in a sandbox's computer use system.

Get API Key

Retrieves an API key by its name.

Get Available Account Providers

Retrieves all available account providers that can be used to authenticate and link accounts.

Get Build Logs URL

Retrieves the build logs URL for a specific sandbox by its ID or name.

Get Commit History

Gets commit history from a repository in a sandbox toolbox.

Get Computer Use Status

Retrieves the status of all VNC desktop processes in a sandbox.

Get Daytona Config

Retrieves the Daytona platform configuration including URLs, OIDC settings, analytics, rate limits, and other platform-level settings.

Get Display Info

Retrieves information about displays connected to a sandbox.

Get Entrypoint Logs

Retrieves entrypoint logs for a sandbox session.

Get File Info (Deprecated)

[DEPRECATED] Gets file info inside a sandbox toolbox.

Get Git History (Deprecated)

[DEPRECATED] Get commit history from a git repository in a sandbox toolbox.

Check Sandbox Access

Checks if the authenticated user has access to a specific sandbox.

Get Daytona Health Status

Retrieves the current health status of the Daytona service.

Get Organization

Retrieves detailed information about an organization by its unique identifier.

Get Organization Audit Logs

Retrieves audit logs for an organization.

Get Organization Invitations Count

Retrieves the count of organization invitations for the authenticated user.

Get Organization Usage Overview

Retrieves the current usage overview for an organization, including CPU, memory, disk, snapshot, and volume usage metrics across all regions.

Get Process Status

Retrieves the status of a specific VNC process in a sandbox.

Get Project Dir (Deprecated)

[DEPRECATED] Gets the project directory path within a sandbox toolbox.

Get PTY Session

Retrieves information about a PTY (pseudo-terminal) session in a sandbox.

Get Push Access

Retrieves temporary storage access credentials for pushing objects to object storage.

Get Docker Registry

Retrieves details of a specific Docker registry by its unique identifier.

Get Sandbox

Retrieves details of a specific sandbox by its ID or name.

Get Session Command Logs

Retrieves logs for a specific command executed within a session.

Get Session (Deprecated)

[DEPRECATED] Retrieves details of a specific session in a sandbox toolbox.

Get Snapshot

Retrieves details of a specific snapshot by its ID or name.

Get Snapshot Build Logs

Retrieves build logs for a specific snapshot.

Get Toolbox Proxy URL

Retrieves the toolbox proxy URL for a sandbox by its ID or name.

Get Transient Push Access

Retrieves temporary registry access credentials for pushing snapshots.

Get Authenticated User

Retrieves the profile of the authenticated user.

Get User Home Dir (Deprecated)

[DEPRECATED] Gets the user home directory path for a sandbox.

Get Volume

Retrieves details of a specific volume by its ID.

Get Windows

Retrieves information about all windows in a sandbox.

Get Working Directory

Gets the current working directory inside a sandbox toolbox.

Get Work Dir (Deprecated)

[DEPRECATED] Retrieves the working directory path of a sandbox toolbox.

List API Keys

Lists all API keys for the organization.

List Available Regions

Lists all available regions for the organization.

List Files (Deprecated)

[DEPRECATED] Lists files and directories in a sandbox toolbox.

List Git Branches (Deprecated)

[DEPRECATED] Lists all git branches from a repository in a sandbox toolbox.

List Organization Invitations

Lists all pending organization invitations.

List Organization Invitations for User

Lists all organization invitations for the authenticated user.

List organization members

Lists all members of a specific organization.

List organization roles

Lists all roles available in an organization.

List Organizations

Lists all organizations that the authenticated user has access to.

List Process Sessions

Lists all active process sessions in a sandbox.

List PTY Sessions

Lists all active PTY sessions in a sandbox.

List registries

Lists all docker registries available in the organization.

List Sandboxes

Lists all sandboxes for the organization.

List Sandboxes Paginated

Lists all sandboxes with pagination support.

List shared regions

Tool to list all shared regions.

List Snapshots

Lists all snapshots for the organization.

List Volumes

Lists all volumes in the organization.

Replace In Files

Replaces text or patterns in multiple files within a sandbox.

Replace Sandbox Labels

Replaces all labels on a sandbox with a new set of key-value pairs.

Resize PTY Session

Resizes an active PTY session with new terminal dimensions.

Revoke SSH Access

Revokes SSH access credentials for a sandbox environment.

Search Files

[DEPRECATED] Search for files inside a sandbox toolbox.

Set Sandbox Auto Archive Interval

Sets the auto-archive interval for a sandbox.

Set Auto Delete Interval

Sets the auto-delete interval for a sandbox.

Set Autostop Interval

Sets the auto-stop interval for a sandbox in the Daytona platform.

Start Computer Use

Starts all VNC desktop processes (Xvfb, xfce4, x11vnc, novnc) for a sandbox.

Start LSP Server

Starts an LSP server process inside a sandbox project.

Start VNC Process

Restarts a specific VNC process within a sandbox in the Daytona platform.

Start Sandbox

Starts a sandbox in the Daytona platform.

Stop LSP Server

Stops an LSP server process inside a sandbox project.

Stop Sandbox

Stops a running sandbox in the Daytona platform.

Take Compressed Screenshot

Takes a compressed screenshot of the entire screen in a sandbox.

Unlink Account

Unlinks a connected provider account (such as GitHub, GitLab, or Bitbucket) from the authenticated user's Daytona profile.

Update Experimental Config

Updates experimental configuration for an organization on the Daytona platform.

Update Last Activity

Updates the last activity timestamp for a sandbox in the Daytona platform.

Update Organization Invitation

Updates an organization invitation with new role and permissions.

Update Organization OpenTelemetry Config

Updates the OpenTelemetry configuration for an organization.

Update Organization Role

Updates an organization role with a new name, description, and permissions.

Update Public Status

Updates the public visibility status of a sandbox in the Daytona platform.

Update Docker Registry

Updates a Docker registry configuration.

Update Sandbox Default Limited Network Egress

Updates the sandbox default limited network egress setting for an organization.

FAQ

Frequently asked questions

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

Yes, you can. Claude Agent 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 Daytona tools.

Yes, absolutely. You can configure which Daytona 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.

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 Daytona data and credentials are handled as safely as possible.

Start with Daytona.It takes 30 seconds.

Managed auth, hosted MCP servers, and every Daytona tool your agent needs.Free to start.

Start building