# Heroku

> Manage Heroku apps, dynos, add-ons, and Postgres from an AI agent.

[Canonical HTML page](https://top-mcps.com/mcp/heroku) · [server.json](https://top-mcps.com/mcp/heroku.json) · [methodology](https://top-mcps.com/about/methodology)

## Install

### Claude Desktop — `claude_desktop_config.json`

Paste under mcpServers. Fully quit and reopen Claude after editing.

```json
{
  "mcpServers": {
    "heroku": {
      "command": "heroku",
      "args": [
        "mcp:start"
      ],
      "env": {
        "HEROKU_API_KEY": "${HEROKU_API_KEY}"
      }
    }
  }
}
```

### Claude Code — `CLI or .mcp.json`

Run from your repo. Commit .mcp.json to share with your team.

```shell
# export HEROKU_API_KEY=your_heroku_token
claude mcp add heroku -- heroku mcp:start
```

### Cursor — `.cursor/mcp.json`

Global path: ~/.cursor/mcp.json. Reload window after editing.

```json
{
  "mcpServers": {
    "heroku": {
      "command": "heroku",
      "args": [
        "mcp:start"
      ],
      "env": {
        "HEROKU_API_KEY": "${HEROKU_API_KEY}"
      }
    }
  }
}
```

### VS Code — `.vscode/mcp.json`

VS Code uses the "servers" key (not "mcpServers").

```jsonc
{
  "servers": {
    "heroku": {
      "command": "heroku",
      "args": [
        "mcp:start"
      ],
      "env": {
        "HEROKU_API_KEY": "${HEROKU_API_KEY}"
      }
    }
  }
}
```

### Windsurf — `~/.codeium/windsurf/mcp_config.json`

Open via Cascade → hammer icon → Configure.

```json
{
  "mcpServers": {
    "heroku": {
      "command": "heroku",
      "args": [
        "mcp:start"
      ],
      "env": {
        "HEROKU_API_KEY": "${HEROKU_API_KEY}"
      }
    }
  }
}
```

### Cline — `cline_mcp_settings.json`

Open via the Cline sidebar → MCP Servers → Edit.

```json
{
  "mcpServers": {
    "heroku": {
      "command": "heroku",
      "args": [
        "mcp:start"
      ],
      "env": {
        "HEROKU_API_KEY": "${HEROKU_API_KEY}"
      }
    }
  }
}
```

### Continue — `~/.continue/config.json`

Continue uses modelContextProtocolServers with a transport block.

```json
{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "stdio",
          "command": "heroku",
          "args": [
            "mcp:start"
          ],
          "env": {
            "HEROKU_API_KEY": "${HEROKU_API_KEY}"
          }
        }
      }
    ]
  }
}
```

### Codex CLI — `~/.codex/config.toml`

Codex uses TOML. Each server is a [mcp_servers.<name>] subtable.

```shell
# ~/.codex/config.toml
[mcp_servers.heroku]
command = "heroku"
args = [
  "mcp:start",
]
env = { HEROKU_API_KEY = "${HEROKU_API_KEY}" }
```

### Zed — `~/.config/zed/settings.json`

Zed calls them "context_servers". Settings live-reload on save.

```jsonc
{
  "context_servers": {
    "heroku": {
      "command": {
        "path": "heroku",
        "args": [
          "mcp:start"
        ]
      },
      "env": {
        "HEROKU_API_KEY": "${HEROKU_API_KEY}"
      }
    }
  }
}
```

### ChatGPT — `ChatGPT → Apps directory`

Heroku doesn't ship a hosted HTTPS endpoint today. ChatGPT supports remote MCP servers only — to use this server in ChatGPT you'll need to deploy it to a public HTTPS URL first (e.g. via Cloudflare Workers or Vercel) or wait for an official remote build.

```none

```

## At a glance

- **Maintainer:** Heroku
- **Transport:** stdio
- **Auth model:** API key
- **Required secrets:** HEROKU_API_KEY
- **Supported clients:** Claude, Cursor, Any MCP-compatible client, Heroku accounts
- **License:** Apache-2.0
- **Language:** TypeScript
- **Latest version:** latest
- **Last verified:** 2026-05-27
- **GitHub stars:** 77 (fetched 2026-06-02T11:55:52.216Z)
- **Score:** 80/100 (rubric 2026-04 — see https://top-mcps.com/about/methodology)
- **Source:** https://github.com/heroku/heroku-mcp-server

## Quick answer

**What it does.** Surfaces Heroku Platform operations — apps, dynos, add-ons, pipelines, Postgres — as MCP tools.

**Best for:**
- App lifecycle (create, rename, transfer)
- Dyno scaling and restart
- Add-on management
- Heroku Postgres ops (info, locks, backups, upgrade)
- Pipeline promote, list, info

**Not for:**
- Unattended production scaling
- Add-on provisioning without cost awareness
- Workflows needing Heroku features outside the platform API

## Description

The official Heroku MCP exposes app lifecycle, dyno scaling, add-on management, and Heroku Postgres operations as MCP tools. The recommended setup uses the existing Heroku CLI authentication via `heroku mcp:start`.

## Why it matters

Heroku is still the default PaaS for thousands of teams. Agent access to apps, dynos, and Postgres collapses common ops chores (restart a dyno, run psql, promote a pipeline) into chat.

## Key features

- App + dyno + add-on tools
- Heroku Postgres ops (psql, info, backups, upgrade)
- Pipeline promote/list/info
- Reuses existing Heroku CLI auth
- Apache-2.0, TypeScript

## FAQ

### Does it require a new API key?

No — the recommended `heroku mcp:start` command uses the existing Heroku CLI auth. The npx variant takes an API key.

### Can it modify production data?

Yes — `pg_psql` runs arbitrary SQL and `ps_scale` modifies live traffic. Keep prod-affecting tools behind a confirmation step.

### What does `MCP_SERVER_REQUEST_TIMEOUT` do?

Sets the request timeout in ms (default 15000). Useful for long Heroku Postgres maintenance ops.

## Changelog

- **2026-05-27** — Refreshed install snippets and fact sheet; verified for 2026.
- **2025-05-10** — Initial directory listing.