Connect your AI tools with MCP

Connect your AI coding tools to Sliplane using the Model Context Protocol (MCP). The Sliplane MCP server is a hosted, remote MCP server available at:

https://mcp.sliplane.io

MCP tools can change or delete resources. Review infrastructure changes before approving them in your client.

What is the Sliplane MCP server?

The MCP server lets AI clients use the Sliplane API through Streamable HTTP.

Authentication

OAuth is the recommended authentication method. When your MCP client connects, it opens Sliplane in your browser so you can sign in and authorize access.

Use these OAuth settings if your client asks for them:

Client ID: sliplane-mcp
Scope: full
Client secret: None

For clients without OAuth support, use a team API key. Keep it out of committed configuration files.

Setup

Choose your MCP client below. The hosted server requires no local installation or self-hosting.

Claude Code

Add the Sliplane MCP server with its OAuth client ID:

claude mcp add --transport http \
  --client-id sliplane-mcp \
  sliplane https://mcp.sliplane.io

Start Claude Code, enter /mcp, and complete the browser authorization flow.

Create an API key, then send it as a bearer token:

claude mcp add --transport http \
  -H "Authorization: Bearer yourapikeyhere" \
  sliplane https://mcp.sliplane.io

Codex

Add the Sliplane MCP server with OAuth:

codex mcp add sliplane \
  --url https://mcp.sliplane.io \
  --oauth-client-id sliplane-mcp

If authorization does not start automatically, run codex mcp login sliplane --scopes full and complete the browser flow. Use a current Codex release that supports --oauth-client-id.

Create an API key, then send it as --bearer-token-env-var:

# Set SLIPLANE_API_KEY privately in the client environment first.
codex mcp add sliplane \
  --url https://mcp.sliplane.io \
  --bearer-token-env-var SLIPLANE_API_KEY

Cursor

Add one of the following configurations to your project-specific or global .cursor/mcp.json file:

.cursor/mcp.json

{
  "mcpServers": {
    "sliplane": {
      "url": "https://mcp.sliplane.io",
      "auth": {
        "CLIENT_ID": "sliplane-mcp",
        "scopes": ["full"]
      }
    }
  }
}

When Cursor connects, complete the OAuth authorization flow.

.cursor/mcp.json

{
  "mcpServers": {
    "sliplane": {
      "url": "https://mcp.sliplane.io",
      "headers": {
        "Authorization": "Bearer yourapikeyhere"
      }
    }
  }
}

opencode

Merge this into your existing configuration:

opencode.json

{
  "mcp": {
    "sliplane": {
      "type": "remote",
      "url": "https://mcp.sliplane.io",
      "oauth": {
        "clientId": "sliplane-mcp",
        "scope": "full"
      }
    }
  }
}

Run opencode mcp auth sliplane and complete browser authorization.

Add this entry to your opencode.json:

opencode.json

{
  "mcp": {
    "sliplane": {
      "type": "remote",
      "url": "https://mcp.sliplane.io",
      "oauth": false,
      "headers": {
        "Authorization": "Bearer {env:SLIPLANE_API_KEY}"
      }
    }
  }
}

Set SLIPLANE_API_KEY privately in the environment that launches opencode.

Visual Studio Code

For an API-key connection, create .vscode/mcp.json. The password input keeps the key out of the configuration file:

.vscode/mcp.json

{
  "inputs": [
    {
      "type": "promptString",
      "id": "sliplane-api-key",
      "description": "Sliplane API key",
      "password": true
    }
  ],
  "servers": {
    "sliplane": {
      "type": "http",
      "url": "https://mcp.sliplane.io",
      "headers": {
        "Authorization": "Bearer ${input:sliplane-api-key}"
      }
    }
  }
}

Start the server in VS Code's MCP controls and enter your key when prompted.

Other MCP clients

For clients not listed above, use the following connection settings:

URL: https://mcp.sliplane.io
Transport: Streamable HTTP
OAuth client ID: sliplane-mcp
OAuth scope: full
OAuth client secret: None

For API-key authentication, send an Authorization: Bearer <API key> header.

Verify the connection

Ask your client to list your Sliplane projects to verify the connection.

Connect your AI tools with MCP

On this page