Deploy from Cursor: connect Livops over MCP
Add Livops to ~/.cursor/mcp.json with just its address, https://livops.uz/api/mcp, and sign in through the browser. Cursor's agent can then deploy, read build logs, explain a failure and roll back, and it asks you before each tool call unless you change that.
By Livops team, the people who build and run LivopsUpdated
Cursor’s agent edits your code; with Livops connected it can also put that code online and tell you why a build broke. This page has the mcp.json entry for both ways of signing in, what each one lets the agent do, and how to keep the agent asking before it changes anything.
What you need
- A Livops account, and a project or a GitHub repository to make one from.
- Cursor.
- For the token route only: an API token from the team’s Settings → API tokens, created by the team owner.
Livops is an MCP server at https://livops.uz/api/mcp, over Streamable HTTP.
Connect through the browser
Put this in ~/.cursor/mcp.json to have Livops in every project, or in .cursor/mcp.json to have it in one[1]:
{
"mcpServers": {
"livops": {
"url": "https://livops.uz/api/mcp"
}
}
}
Cursor supports OAuth for servers that require it, with the authorization code flow and PKCE[1]. On first connect Livops opens in the browser: you see the app’s name, pick a team, see what it will be able to do, and press Allow. With two-factor authentication on, you enter your code first.
Cursor gets a key that lasts an hour, works only at /api/mcp and renews itself. The connection shows under Account → Connected AI apps; Disconnect stops it at once.
Connect with an API token
Cursor resolves variables in url and headers, including ${env:NAME}[1], so the token can stay out of the file:
{
"mcpServers": {
"livops": {
"url": "https://livops.uz/api/mcp",
"headers": { "Authorization": "Bearer ${env:LIVOPS_TOKEN}" }
}
}
}
Set LIVOPS_TOKEN to an lvp_ token in the environment Cursor starts from. This matters most for .cursor/mcp.json, which sits in the project and is easy to commit. Give Cursor its own token, so revoking it stops nothing else.
What can Cursor’s agent do in Livops?
It works in one team, as you, with your role in that team. A viewer only reads.
| Browser sign-in | lvp_ API token |
|
|---|---|---|
| Read projects, deploys, build and runtime logs, metrics | Yes | Yes |
| Deploy, roll back, restart, set variables | Yes | Yes |
| Delete anything | Yes, if your role allows | No |
| Read a database password or bucket key | Yes, if your role allows | No |
| Open a Payme or Click top-up link, redeem a promo code | Yes | No |
There is no tool for card numbers, your password, two-factor, sessions, creating API tokens, deleting the account, handing a project over, project members or downloading a raw backup. Those stay in the browser.
Questions that work well:
- “Why did the last deploy of shop-api fail?” (
get_deploy_log,explain_failure) - “Which project used the most memory in the last 24 hours?” (
get_metrics) - “Roll shop-web back to the previous version.” (
rollback) - “Which of my projects are down right now?” (
list_projects)
A project connected to a GitHub branch deploys on every push to it, so the agent can push, then follow the build with get_deploy_log. If the build fails, the previous version keeps running.
Keep the confirmations on
Cursor asks for approval before it uses an MCP tool by default, and you can open the tool’s arguments before you approve[1]. MCP calls follow the same Run Modes as terminal commands; in Auto-review, allowlisted MCP tools run immediately and everything else goes through a classifier[1].
Livops marks every read-only tool with readOnlyHint and every tool that changes something with destructiveHint. The specification says a human should be able to deny a tool call[2], and it also says a client must treat annotations as untrusted unless the server is trusted[2]: the hint helps you decide, but it is your approval that holds. A reasonable split is to allowlist list_* and get_* and approve the rest one at a time.
Logs and commit messages reach the model as data. A build log that contains “ignore previous instructions” is still a log, but the agent decides what to do with it, and your approval is the step after that.
Deploying from Cursor on Livops
- Add the
livopsentry to~/.cursor/mcp.jsonwith just the URL. - Let Cursor open the browser, pick the team, press Allow.
- In the agent, ask: “List my Livops projects.”
- Ask it to deploy, or push and ask it to watch the build.
The full tool list is in MCP (AI clients); what a token may do is in API tokens.
Questions
Where does Cursor keep the MCP config?
In ~/.cursor/mcp.json for tools available in every project, or .cursor/mcp.json inside one project.
Do I need an API token for Cursor?
No. With just the address, Cursor signs you in through the browser, and you pick a team and press Allow in Livops. A token is for machines without a browser and for teams that prefer a key they revoke by hand.
Will Cursor's agent deploy without asking me?
Not by default: Cursor asks for approval before it uses an MCP tool. In Auto-review mode, tools you allowlisted run at once and the rest go through Cursor's classifier, so allowlist only the tools that read.
Can Cursor delete my project?
Only when it is connected through the browser and your role in the team allows deleting. With an lvp_ API token every delete is refused by the server, whatever the agent asks for.
Does Cursor see my secrets?
Not the variables: list_variables returns names and which are secret, never values. A browser connection can ask for a database connection URL or a bucket key, which are real secrets, so do not approve those calls in a chat you share.
How do I disconnect Cursor?
In the cabinet, Account, then Connected AI apps, then Disconnect. If you used a token, revoke it under Settings, API tokens. Either one stops Cursor at once.
Sources
- Cursor — Model Context Protocol (MCP) (cursor.com). Checked .
- Model Context Protocol — Tools (specification 2025-11-25) (modelcontextprotocol.io). Checked .
Put your project online today
Connect GitHub or drop a folder. No card is asked for.
Connect CursorRelated
- MCP server for deploying apps: what to look forAn MCP deploy server lets Claude Code, Cursor, Codex or Copilot deploy, read logs and roll back on your hosting account. Judge one by how it signs in, how far a key reaches, whether its tools say which ones change things, and what you can see and revoke afterwards. Vercel, Render, Railway, Netlify and Livops all ship one.
- Deploy from Claude Code: connect Livops over MCPRun one command, claude mcp add --transport http livops https://livops.uz/api/mcp, then sign in through /mcp. From then on Claude Code can list your projects, deploy, read the build log, explain a failure and roll back, as you and only in the team you picked.
- Deploy from VS Code: Copilot agent sessions and the Livops MCP serverIn VS Code, add Livops to the workspace's .mcp.json as an http server at https://livops.uz/api/mcp, select the Copilot harness in a chat session, and complete the sign-in when prompted. Copilot can then deploy, read build logs, explain a failure and roll back, and each tool call still needs your approval.
- What Livops's AI assistant does when a build failsWhen a build fails on Livops, the previous version keeps running and one button asks the assistant why. It reads the failed build's log and your project's commands, says what went wrong and how to fix it, and may offer one fix as a button that changes nothing until you press it. It never deletes, never sees variable values, and does not edit your code.