# Daytona MCP for AI Agents AI Agent Connect > Daytona MCP for AI Agents. Manage your cloud-based dev environments and sandboxes directly through your AI agent. Create, scale, and snapshot ephemeral workspaces with specific CPU and RAM configs. Stop manual environment setup and let your agent handle the infrastructure lifecycle. ## Overview - **Category:** developer-tools - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_AkqYHK4XFPqtgrdtnVnobFEfCwP5QQF4omfr3EXo/ai-agent-connect - **Tags:** sandboxes, dev-environments, workspace-automation, ephemeral-infrastructure, cloud-development ## Description Managing dev environments is usually a mess of manual clicks and configuration drift. You're constantly spinning up boxes, hoping the dependencies are right, and then worrying about the costs of leaving them running. This Connector changes that by giving your agent the keys to your Daytona account. Instead of jumping between the CLI and a dashboard, you just tell your agent what you need. It handles the heavy lifting of provisioning sandboxes with the exact specs you require. If you need to scale up your RAM because a build is hanging, or if you need to grab a snapshot of a working environment to fork for a new feature, you just ask. It makes the whole infrastructure feel like a part of your chat interface. By connecting this through the Vinkius catalog, you're moving from manual environment management to a conversation-driven workflow where your dev space is ready as soon as you're ready to code. For QA teams, it means getting a fresh environment for a bug report in seconds. For DevOps, it means ensuring every sandbox is provisioned with the exact same configuration every single time. It turns environment management from a chore into a quick command. ## Tools ### archive_sandbox Move a sandbox into an archived state for long-term storage. Keep your active list clean while saving work. ### list_snapshots View all the snapshots currently saved in your account. See your history of saved states. ### create_api_key Generate a new API key for your Daytona account. Manage your access tokens directly through your agent. ### create_snapshot Save the current state of a sandbox into a new snapshot. It's great for capturing a stable point in your dev work. ### create_volume Create a new persistent volume for your dev environments. Add dedicated storage for your project data. ### deactivate_snapshot Turn off a specific snapshot so it's no longer active. Use this to manage your snapshot lifecycle. ### delete_api_key Remove an old API key by its name to keep your account clean. It helps maintain tight security. ### delete_sandbox Permanently remove a sandbox you no longer need. This helps keep your workspace organized. ### delete_snapshot Delete a snapshot to free up storage space. Keep your saved states lean and relevant. ### delete_volume Remove a persistent volume from your account. Use this when you no longer need associated data. ### fork_sandbox Create a new sandbox that's a copy of an existing one. This is perfect for starting new features from a known state. ### get_api_key Fetch the specific details of an API key you already have. Quickly check your credentials in one place. ### get_sandbox View all the current specs and details for a specific sandbox. Check your active resources at a glance. ### get_snapshot Retrieve the details of a specific snapshot. See what's saved in your environment history. ### get_volume_by_name Find the details of a volume using its specific name. Easily locate persistent storage by its label. ### get_volume Get the details for a volume using its unique ID. Retrieve storage info for specific environment components. ### list_api_keys See a full list of all API keys in your Daytona account. Audit your access points in seconds. ### list_sandboxes_paginated View your sandboxes in a paginated list for large accounts. Handle many environments without clutter. ### list_sandboxes See a complete list of all your current sandboxes. Get a high level view of your dev workspace. ### list_volumes See a list of all persistent volumes you have active. Track your storage usage across environments. ### recover_sandbox Fix a sandbox that's stuck in an error state. Get back to coding without a full reinstall. ### resize_sandbox Change the CPU, RAM, or Disk size of an active sandbox. Scale your resources up or down instantly. ### start_sandbox Turn on a sandbox that is currently stopped. Resume your work with a single command. ### stop_sandbox Shut down a running sandbox to save on resources. Turn off environments when you're done for the day. ### create_sandbox Create a new Daytona sandbox with your preferred hardware specs. This lets you spin up fresh environments instantly. ### get_current_api_key See the details of the key currently being used for authentication. Useful for checking active sessions. ### get_sandbox_preview_url Get a signed URL to view a specific port on a sandbox. Share your work with teammates easily. ### activate_snapshot Turn a saved snapshot into a live, active sandbox. Quickly spin up a previous state as a new environment. ## Prompt Examples **Prompt:** ``` List my sandboxes and tell me which ones are stopped. ``` **Response:** ``` I've pulled your current Daytona sandboxes. Here's the status: | Sandbox Name | Status | Resources | | :--- | :--- | :--- | | **web-app-dev** | Active | 2 vCPU, 4GB RAM | | **api-testing** | **Stopped** | 1 vCPU, 2GB RAM | | **db-sandbox** | **Stopped** | 4 vCPU, 8GB RAM | Would you like me to start any of the stopped ones? ``` **Prompt:** ``` I need a new sandbox for a Node.js app with 4GB of RAM. ``` **Response:** ``` I'm starting the creation of a new Node.js sandbox for you. I'll set it up with 4GB of RAM and let you know as soon as it's ready. **Provisioning Details:** - Image: `node:20` - Memory: 4GB - Status: Provisioning... ``` **Prompt:** ``` The 'web-app' sandbox is crashing. Can you fix it? ``` **Response:** ``` I'll check that for you. I'm running the recovery tool on the 'web-app' sandbox now. **Status Update:** - Action: `recover_sandbox` - Target: `web-app` - Result: **Success.** The environment has been reset to its last stable state. You should be able to resume work now. ``` ## Capabilities ### Provision new sandboxes Create a fresh workspace with specific hardware specs. ### Scale environment resources Adjust CPU, RAM, or Disk on the fly. ### Capture environment states Create snapshots to save your current progress. ### Fork existing workspaces Duplicate a sandbox to start a new task from a known state. ### Recover crashed environments Restore a sandbox to a functional state automatically. ### Manage access tokens List and create API keys to keep your account secure. ## Use Cases ### Branch Isolation A developer needs a clean environment for a new feature. They ask the agent to `fork_sandbox` from the main dev environment to keep things separate. ### Resource Bottlenecks A build is failing due to low memory. The user tells the agent to `resize_sandbox` to give it 16GB of RAM, then scales it back once finished. ### QA Replication A QA tester finds a bug and wants to share the exact environment. They use `create_snapshot` and share the snapshot ID with the team. ### Automated Cleanup A DevOps engineer asks the agent to `list_sandboxes` and `delete_sandbox` for any that haven't been used in a week to keep costs down. ## Benefits - Stop manual provisioning. Use `create_sandbox` to get a ready-to-code environment in seconds without touching a dashboard. - Scale on demand. Use `resize_sandbox` to bump your RAM or CPU when a heavy build starts, then scale back down. - Preserve progress. Use `create_snapshot` to save your setup so you can return to it later or fork it for other tasks. - Faster debugging. Use `recover_sandbox` to instantly fix environments that crash, saving hours of manual reconfiguration. - Secure access. Use `list_api_keys` and `create_api_key` to manage your credentials without leaving your agent's chat. ## How It Works The bottom line is you stop clicking buttons to manage infrastructure and start describing the environments you need. 1. Subscribe to the Daytona MCP and provide your API key. 2. Connect it to your favorite client like Claude or Cursor. 3. Ask your agent to spin up, scale, or snapshot your dev environments. ## Frequently Asked Questions **What does the Daytona MCP do for my team?** It gives your team the ability to manage cloud dev environments using natural language. You can create, scale, and snapshot sandboxes without leaving your preferred AI client. **Can I use it to manage my cloud dev environments?** Yes. This Connector connects your agent to Daytona so you can handle the entire lifecycle of your cloud-based dev workspaces through conversation. **How does it help with sandbox management?** It automates the tedious parts of sandbox management. You can fork existing environments, recover crashed ones, and manage snapshots to keep your workspace organized. **Can the agent scale my resources automatically?** Absolutely. You can just tell your agent to give a sandbox more RAM or CPU, and it will execute the resize command for you instantly. **Is it safe to manage API keys through this?** Yes, it provides tools to list and create API keys directly. This allows you to manage your Daytona account credentials securely within your agent's workflow. **How do I recover a crashed environment?** You can simply ask your agent to recover a specific sandbox. It will trigger the recovery process to bring the environment back to a functional state. **Can I change the resources of an existing sandbox?** Yes, you can use the `resize_sandbox` tool to update the CPU, memory, or disk size of a specific sandbox by providing its ID or name. **How do I create a new environment with a specific Docker image?** Use the `create_sandbox` tool and provide the `image` parameter with the desired Docker or OCI image name. **What should I do if a sandbox is in an error state?** You can use the `recover_sandbox` tool to attempt to restore a sandbox from an error state back to a functional one.