Heroku — Cursor MCP server: what it does, how to install
MCP server connecting AI coding assistants directly to Heroku platform resources and environments.
no license updated 2026-10-29
Open on GitHubWhat Heroku does
This server connects AI editors to the Heroku platform through the Model Context Protocol. It allows assistants to inspect, configure, and operate resources across Common Runtime, Fir Private Spaces, and Cedar Private or Shield Spaces.
Two startup options exist: invoking heroku mcp:start or executing npx -y @heroku/mcp-server. The first route is cleaner because it reuses existing Heroku CLI credentials, avoiding manual secret handling.
A global installation of Heroku CLI version 10.8.1 or newer is mandatory. The server fails immediately if the CLI binary is missing or outdated.
What it helps with
- inspect Heroku application settings through assistant prompts
- manage deployment environments across Common Runtime and Shield Spaces
- authenticate AI tools via Heroku CLI sessions without sharing static tokens
Example prompts for the agent
- List the Heroku apps currently running in my account
- Check the dyno formation and runtime configuration on Heroku
- Inspect environment details for the app in Fir Private Spaces
How to install Heroku
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Heroku
How do I configure the Heroku MCP server in an MCP client?
Install Heroku CLI v10.8.1+ globally and set heroku mcp:start as the executable command inside your client MCP configuration block.
What does the Heroku MCP server do?
It exposes Heroku platform APIs to language models so they can query and operate cloud resources using conversational instructions.
Do I need to generate a HEROKU_API_KEY?
Only if you launch with npx -y @heroku/mcp-server. The recommended heroku mcp:start workflow inherits active session credentials from your Heroku CLI automatically.