Cursor MCP Server
Connect to the Cursor MCP server to launch, monitor, and control cloud coding agents across your GitHub repositories using AI agents on Gumloop, Claude, or Cursor.
Try Cursor now
Type what you want done. Sign in and run it live with an AI agent.
|
Cursor
Installation
Get StartedCreate a Gumloop Account
To use this MCP, you need a Gumloop account. If you don't have one yet, you can create one for free.
Copy Your Server URL
Copy your MCP server URL and add it to your client. You'll be prompted to authorize on first use.
Tools (9)
List Agents
List all cloud agents for the authenticated user
Launch Agent
Launch a new cloud agent to work on a GitHub repository
Get Agent
Get a cloud agent's status and results
Get Agent Conversation
Get a cloud agent's conversation history
Add Followup
Send a follow-up instruction to a running cloud agent
Stop Agent
Stop a running cloud agent
Delete Agent
Permanently delete a cloud agent
List Models
List available LLM models for cloud agents
List Repositories
List accessible GitHub repositories
What is Cursor MCP?
The Cursor MCP server gives AI agents access to Cursor’s cloud coding agents, the headless background agents that work on your GitHub repositories. That means an AI agent can launch a new cloud agent against a repo, list the agents you already have running, check an agent’s status and results, read its full conversation history, send follow-up instructions while it works, and stop or delete an agent when it is done. It can also list the models available for a run and list the repositories your account can reach. It works across the full cloud-agent lifecycle: launch, monitor, follow up, and shut down.
If your team kicks off Cursor background agents one at a time, tabs back and forth to check whether each run finished, and copies status updates into Slack or a tracker by hand, an AI agent can take over that coordination. Describe what you need, and your AI agent will launch and manage the cloud agents for you.
MCP stands for Model Context Protocol. It is an open standard that gives AI agents a way to connect to external tools and services. Instead of writing code against the Cursor Cloud Agents API, handling Basic Auth, polling for status, and paging through results yourself, you add your Cursor API key to Gumloop once. After that, you can launch and orchestrate cloud coding agents just by chatting with your AI agent.
Related MCP servers
Devin MCP Server
Interact with Devin AI coding sessions
Github MCP Server
Server for interacting with Github
Linear MCP Server
Interact with Linear issue tracking and project management
JIRA MCP Server
Interact with JIRA for project management, issue tracking, and user information access.
Slack MCP Server
Interact with Slack channels and messages
Notion MCP Server
Server for interacting with Notion
What you can do with Cursor MCP on Gumloop
Launch cloud coding agents on a repository
Start a new Cursor cloud agent against a GitHub repo with a prompt describing the task. The agent writes code on a branch and can automatically open a pull request when it finishes, so an AI agent can kick off work without anyone opening Cursor.
List and monitor your agents
Pull every cloud agent for your account, then check any single agent’s status and results as it moves from creating to running to finished. Your AI agent can watch a batch of runs and report back only when something needs attention.
Send follow-up instructions mid-run
Add a follow-up prompt to a running agent to refine the task, correct course, or add scope. The agent picks up the new instruction without you starting over.
Stop or delete agents
Stop a running agent when a task is no longer needed, or permanently delete an agent to clean up. Deletion is irreversible, so an AI agent can be told to delete only after you confirm.
Read an agent’s full conversation history
Retrieve the complete prompt-and-response thread for any agent. Useful for auditing what an agent was asked to do, debriefing a run, or summarizing the work into a ticket or release note.
Discover available models
List the LLM models available for cloud agent runs so an AI agent can pick the right model for a given task before it launches one.
List accessible GitHub repositories
Enumerate the repositories your Cursor account can reach, so an AI agent can confirm a target repo exists before launching an agent against it.
How to connect the Gumloop Cursor MCP Server
- 1
Create a free Gumloop account
Sign up at gumloop.com. No credit card required.
- 2
Add the Cursor MCP server
Copy your MCP server URL from Gumloop and add it to your preferred client (Claude, Cursor, or Gumloop workflows). You'll authorize on first use.
- 3
Start using Cursor in your AI workflows
That's it. Your AI agent can now launch, monitor, follow up on, and stop cloud coding agents across your repositories. Use it inside a Gumloop automation, in Claude Desktop, or in Cursor.
Cursor MCP use cases
Backlog automation for engineering managers
When a sprint has a stack of small, well-scoped tickets, a Gumloop agent can read each one, launch a Cursor cloud agent against the right repo with the ticket as the prompt, and post the resulting pull request links back to the team in Slack. Managers clear routine work without assigning every ticket by hand.
Bug triage to fix for development teams
When a new bug report lands, an AI agent can launch a Cursor cloud agent on the affected repo with a reproduction and a fix prompt, monitor the run to completion, and comment the opened PR on the original issue. The team reviews a proposed fix instead of starting from a blank editor.
Agent fleet monitoring for platform and DevEx teams
For teams running many cloud agents at once, an AI agent can list every active agent, check each one’s status, and roll up a single summary of what finished, what is still running, and what failed. Drop that summary into Slack or a Google Sheet so nobody has to poll Cursor manually.
Run auditing and debriefs for tech leads
After agents finish, an AI agent can pull each one’s full conversation history, summarize what was asked and what changed, and compile a debrief into a Notion page or a Linear comment. Leads get a clear record of what the cloud agents did without reading raw logs.
Incident-driven remediation for on-call engineers
Connect Cursor with Slack, GitHub, and an incident tool in a single agent. When an incident is declared, an AI agent can launch a cloud agent to draft a fix on the relevant repo, send a follow-up if the on-call engineer adds detail, and stop the agent once the human takes over. Cursor’s cloud agents talk to the rest of your stack automatically.
Why use Gumloop for Cursor MCP
Your API key, stored securely, no env vars and no code
Most Cursor MCP setups you’ll find on GitHub require you to keep your Cursor API key in environment variables and write code to handle Basic Auth, polling, and pagination against the Cloud Agents API. With Gumloop you add your key once and it is stored securely, with no config files, no token management, and no code.
Works with multiple MCP clients
Use the Cursor MCP server endpoint in Claude Desktop, Cursor, or directly inside Gumloop agents. Same server URL, works with any MCP client.
Chain Cursor with 100+ other integrations
Combine Cursor with GitHub, Slack, Linear, Jira, Google Sheets, and other MCP tools in a single AI agent. An agent can read a ticket from Linear, launch a Cursor cloud agent to address it, and post the resulting PR to Slack.
Enterprise-grade and scalable
Built for teams, with role-based permissions and dedicated support for Pro users. For details on Gumloop’s security practices, see trust.gumloop.com.
Pricing includes a free plan
You can test the Cursor MCP integration on Gumloop’s free tier before committing. Paid plans start at $37/month.
Frequently asked questions
Ship Cursor agents in minutes
Connect any AI agent to 100+ MCP servers, zero setup.