> Tools for working with Runsite outside the dashboard: the runsite command-line interface for your terminal and CI, and an MCP server for AI coding agents.

# Runsite outside the dashboard

The dashboard is one way to run your services. The CLI and the MCP server reach the same API from a terminal, a build pipeline or an AI agent.

[Install the CLI](/tools/cli)[Connect an agent](/tools/mcp)

bash

 git push origin main

 runsite deploy api

 building 8e4c1a2 on main .......... 24.1s

 health check /healthz ........... 200 OK

 released in 38s

Start the deploy wherever you are. The release comes out the same.

[ Command linerunsite v0.1.0Deploy a service, read its logs and change its environment from your shell. It is one binary for macOS, Linux and Windows, and your build pipeline runs the same one.Runs onmacOS, Linux, Windows (5 builds)Signs in withrunsite login, or RUNSITE\_API\_TOKEN in CIOutput\--output json on anything that prints dataInstallChecked against the release's SHA‑256 sums runsite deploy apiInstall and use it](/tools/cli)[ MCP servermcp v0.1.0Point Claude Code, Cursor or another MCP client at one hosted endpoint. The agent can look at your services, follow a failing build and deploy the fix, with a key whose scope you choose.Runs onOur servers, so there is nothing to installOffers22 tools, filtered by the key's scopeLimit60 tool calls a minute per keyNever returnsEnv values, secrets or connection URLshttps://mcp.runsite.app/mcpConnect a client](/tools/mcp)

## Which one for which job

Most everyday jobs work from all three. The table gives the command or tool name for each one and shows where you still need the dashboard.

| Job                         | Dashboard     | CLI                                       | MCP server                                  |
| --------------------------- | ------------- | ----------------------------------------- | ------------------------------------------- |
| Deploy the latest commit    | Yes           | runsite deploy api                        | deploy                                      |
| Read recent logs            | Yes           | runsite logs api                          | get\_logs                                   |
| Restart or stop a service   | Yes           | runsite service restart api               | restart, stop                               |
| Set an environment variable | Yes           | runsite env set api KEY=value             | set\_envYou type the value in the dashboard |
| Find out why a build failed | Yes           | Not available                             | get\_deployment                             |
| Roll back a deployment      | Yes           | Not available                             | rollbackYou approve it in the dashboard     |
| Deploy from a CI job        | Not available | RUNSITE\_API\_TOKENNo interactive sign-in | Not available                               |
| Open a shell in a container | Yes           | Not yet                                   | Not available                               |

## One API key works for both

Create a key under Settings → API keys and give it a scope. Read lets it look around, write lets it deploy and restart, and admin adds requests for secrets and deletions. Revoke the key in the dashboard and both tools lose access on their next call.

[Create an API key](https://dashboard.runsite.app/settings/api-keys)

`ak_live_…`readwriteadmin

 The CLI reads it from the environment

` export RUNSITE_API_TOKEN=ak_live_…`

 An MCP client sends it with every call

`Authorization: Bearer ak_live_…`

## Your app deserves to be online

€5 of credit on signup. Deploy in under a minute. No credit card needed.

[Start deploying](https://dashboard.runsite.app/login)[View documentation](https://docs.runsite.app)

---

Source: https://runsite.app/tools