Skip to content

Connect an AI IDE or agent

livemy.app runs a remote MCP (Model Context Protocol) server, so you can connect your AI editor or coding agent — Cursor, Claude Code, Windsurf, Claude Desktop, ChatGPT, or any other MCP client — to your livemy.app account and work with it from inside the tool. MCP is an open standard that lets AI tools talk to external services; once connected, your agent acts on your account with the same permissions you have in the dashboard.

Open the account menu and choose Developer console, or go to Settings → Developer console. Both open the MCP Integration page, where you manage access and copy ready-made setup snippets.

Connecting an MCP client requires a paid plan. Any signed-in account can open the page, but if your plan doesn’t include MCP you’ll see an upgrade prompt in place of the connection tools — see the Billing page in your dashboard for what your plan includes.

There are two ways to connect:

  • A personal access token — create a token and paste a config snippet into your client. This is how you connect code editors like Cursor, Claude Code, and Windsurf, and it works for any MCP client.
  • Signing in — add livemy.app as a connector in Claude or ChatGPT and authorize by signing in, with no token to manage.

Both methods use one address — your Server URL. The MCP Integration page shows it with a Copy button. On production it is:

https://my.livemy.app/api/mcp/messages

Copy the value shown on your page rather than typing it — it’s the same address for every method and every client.

Best for code editors and any other MCP client. You create a token, then paste a ready-made config snippet into your client.

  1. On the MCP Integration page, under Active tokens, select New token (or Create your first token if you don’t have any yet).

  2. Give it a name you’ll recognize later, for example “Cursor on my laptop”.

  3. Choose when it expires — 30 days, 90 days (recommended), 1 year, or never. A shorter lifetime is safer, and you can always create a new one.

  4. Select Generate token.

Your token is shown once, in a highlighted card. Copy it immediately and store it somewhere safe — livemy.app never shows it again. If you lose it, revoke it and create a new one.

After you create a token, the Connect a client section pre-fills its snippets with your new token. Pick your client’s tab, copy the snippet, and paste it into the client’s config file. If you come back later, the snippets show a <YOUR_TOKEN> placeholder instead — replace it with your token.

Config file: .mcp.json in your project root (recommended), or ~/.claude.json to make it available in every project.

{
"mcpServers": {
"livemy": {
"type": "http",
"url": "https://my.livemy.app/api/mcp/messages",
"headers": {
"Authorization": "Bearer <YOUR_TOKEN>"
}
}
}
}

Add the livemy entry, keeping any other servers you already have. Run claude from that folder and check the connection with the /mcp command.

If you use an agentic editor like Claude Code, Cursor, or Windsurf, you can skip the manual steps. After you create a token, copy the agent prompt from the page and paste it into your editor’s chat. The agent detects which client you’re using, writes the config file for you, and tells you the next step. It writes the token straight into the config and won’t repeat it back in the chat.

Claude and ChatGPT can connect without a token. You add livemy.app as a custom connector and authorize by signing in.

  1. In Claude (claude.ai or Claude Desktop) or ChatGPT, open Settings → Connectors and add a custom connector.

  2. Paste your Server URL as the connector’s address.

  3. Select Connect. livemy.app opens a sign-in page.

  4. Sign in with your livemy.app account — or Continue as your signed-in account — to authorize access. Signing in is the approval; there’s no separate consent step.

Once connected, the app shows up under Active connections on the MCP Integration page, where you can disconnect it at any time.

The MCP Integration page keeps two separate lists.

  • Active tokens — your personal access tokens. Each shows its name, a short prefix (never the full token), when it expires, and when it was last used. To remove one, select the delete icon and confirm. Revoking is permanent and immediately cuts off any client using that token. You can hold a limited number of active tokens at once; the page shows your current count.
  • Active connections — apps you connected by signing in. Each shows the app and when it connected. Select Disconnect to revoke its access immediately; the app has to reconnect to regain it. This list appears once you have at least one such connection.

Once connected, your AI tool works through the same permissions you have in the dashboard. It can browse your projects, read their deploy reports and AI Assist history, and — depending on your plan — make changes such as deploying and redeploying projects, managing custom domains, chatting with AI Assist, and running commands or editing files on a project’s container. Which actions are available follows your plan; see the Billing page in your dashboard for what’s included.

  • Lost your token? You can’t view it again. Revoke it and create a new one.
  • Client can’t authenticate? Check that the token was copied in full and hasn’t expired, and that you’re using the Server URL shown on your page.
  • “livemy” doesn’t show up in your client? Restart or reload the client after editing its config — many clients only read MCP settings at startup.
  • Claude Desktop won’t start or hangs? Pre-install the bridge (npm i -g mcp-remote) and make sure Node.js is on your PATH. Use the bridge snippet — the plain url form doesn’t work in Claude Desktop.
  • SSH access — open a shell on your project (Pro plan).
  • Using AI Assist — make changes by chatting with the agent in the dashboard.