MCP server

slipway ships a first-party Model Context Protocol server, so AI coding agents (Claude Code, Cursor, Claude Desktop, and anything else that speaks MCP) can drive your environments directly. An agent can bring up an isolated environment from a branch, get a live URL, read logs, run commands to verify its work, check how much capacity you have left, then tear it down. It is the verification loop agents otherwise lack. On the hosted platform the endpoint is https://mcp.slipway.sh/mcp. Settings → MCP in the console shows the exact endpoint for your installation.

Create a token

In the console, go to Settings → MCP (admin only) and choose New token. Give it a name, pick a role, and optionally an expiry. The token is shown once, right after you create it, so copy it then. If you lose it, revoke it and make a new one. Tokens are scoped to a single organization and can be revoked at any time from the same page.

Connect Claude Code

claude mcp add --transport http slipway https://mcp.slipway.sh/mcp \
  --header "Authorization: Bearer YOUR_TOKEN"

Connect Cursor

Add the server to .cursor/mcp.json (in the project, or globally in ~/.cursor/mcp.json):

{
  "mcpServers": {
    "slipway": {
      "url": "https://mcp.slipway.sh/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Connect Claude Desktop

Add an HTTP MCP server with the same URL, https://mcp.slipway.sh/mcp, and an Authorization: Bearer YOUR_TOKEN header. The same shape works for any other MCP client that supports HTTP servers with custom headers.

Roles

A token’s role caps what its agent can do. Each tool requires at least the role shown.

Role	Can do
viewer	Read environments, deployments, logs, repositories, and usage.
developer	Everything a viewer can, plus bring up instances, deploy/redeploy, cancel, tear down, and run commands in a container.
admin	Everything, plus writing variables and secrets.

Give an agent the lowest role that lets it do its job. A read-only agent that just inspects builds only needs viewer; for a coding agent, developer is the usual choice and the default.

The tools

Sixteen tools, listed with the minimum role each requires.

Tool	What it does	Role
`get_usage`	Plan and usage: environments, concurrent deployments, seats, storage, each used vs. cap.	viewer
`check_capacity`	Preflight: can the org start another deployment right now? Returns the blocking limit if not.	viewer
`list_environments`	List base environments with id, name, kind, and aggregate status.	viewer
`get_environment`	One environment’s status and the components that make up its stack.	viewer
`get_deployment`	A deployment’s status, phase, failure reason, and public URLs.	viewer
`list_repos`	The repositories slipway knows about, with default branch and enablement.	viewer
`get_deployment_logs`	Build and runtime logs, optionally filtered to one service and tailed to the last N lines.	viewer
`create_environment`	Create a base environment for a repository and deploy it, permanent or ephemeral.	developer
`bring_up_instance`	Bring up a fresh, isolated ephemeral instance of an existing environment at a branch.	developer
`wait_for_deployment`	Block until a deployment reaches a terminal state or the timeout elapses.	developer
`redeploy_deployment`	Re-run an existing deployment at the same commit.	developer
`cancel_deployment`	Cancel an in-flight deployment.	developer
`teardown_environment`	Tear down an environment or instance. Deployment history is retained.	developer
`exec_in_service`	Run a command in a running service’s container; captures stdout, stderr, and the exit code.	developer
`set_variable`	Set a plaintext variable at org or environment scope.	admin
`set_secret`	Set an encrypted secret at org or environment scope. Write-only: no tool returns a secret value.	admin

Agents are told to check get_usage / check_capacity before bringing things up, so they don’t hit your concurrency cap. Instances an agent brings up are ephemeral by default and auto-expire after a TTL, so a forgotten one cleans itself up. For the design story behind the tools, and a worked end-to-end session, see How to let a coding agent deploy and test its own code.

Security

Tokens are stored hashed; the plaintext only exists at create time.
Every tool is scoped to the token’s organization — an agent can never reach another org’s data.
The exec and secret-writing tools are powerful (arbitrary commands in your pods; secret values). Keep tokens to the minimum role, prefer short expiries for shared machines, and revoke anything you’re unsure about.

Get started

Slipway Spec

Deployments

Environments

Configuration

Clusters

Organization

CLI

Reference

Create a token

Connect Claude Code

Connect Cursor

Connect Claude Desktop

Roles

The tools

Security

​Create a token

​Connect Claude Code

​Connect Cursor

​Connect Claude Desktop

​Roles

​The tools

​Security

Create a token

Connect Claude Code

Connect Cursor

Connect Claude Desktop

Roles

The tools

Security