MCP Server

What it is

The Model Context Protocol (MCP) is the open standard AI assistants use to call tools in other products. SiteLift exposes 44 tools and 3 ready-made prompts over MCP, so instead of opening the dashboard you can ask things like "which of my articles sit on page two and deserve a refresh?", "does ChatGPT recommend us yet?" or "publish the pricing comparison" and let the assistant do the clicking.

A connection acts as you: every site you own or are a member of, with your role on each. Tools go through the same access checks and plan entitlements as the dashboard, so an agent can never do something you could not do yourself. Tools that change your live site are flagged so the assistant confirms with you before calling them.

What you can ask

Built-in prompts

Clients that list MCP prompts (Claude, Claude Code, Cursor and others) show three SiteLift workflows as slash commands:

• /weekly_report: what changed this week across traffic, articles, refreshes and visibility.
• /quick_wins: pages ranking just off page one, with the refresh or edit that would move them.
• /review_calendar: walk the upcoming calendar and tighten briefs and priorities.

Connect a chat assistant

Chat assistants sign in with your SiteLift account. No key to copy: you approve the connection once on a SiteLift consent page and can disconnect it from the dashboard later.

Connect a coding agent

Claude Code

Run once in your terminal, then type /mcp inside Claude Code to confirm SiteLift is connected. Without the header line, Claude Code signs in through the browser instead.

claude mcp add --transport http sitelift https://api.sitelift.io/mcp \
  --header "Authorization: Bearer slk_your_api_key_here"

Codex CLI

Export the key in your shell profile so Codex can read it, then add the server and restart Codex.

# in your shell profile: export SITELIFT_API_KEY=slk_your_api_key_here

[mcp_servers.sitelift]
url = "https://api.sitelift.io/mcp"
bearer_token_env_var = "SITELIFT_API_KEY"

Gemini CLI

{
  "mcpServers": {
    "sitelift": {
      "httpUrl": "https://api.sitelift.io/mcp",
      "headers": { "Authorization": "Bearer slk_your_api_key_here" }
    }
  }
}

Cursor

Use ~/.cursor/mcp.json for every project or .cursor/mcp.json inside one. Cursor Settings → MCP shows a green dot once the tools loaded.

{
  "mcpServers": {
    "sitelift": {
      "url": "https://api.sitelift.io/mcp",
      "headers": { "Authorization": "Bearer slk_your_api_key_here" }
    }
  }
}

VS Code

Copilot agent mode reads .vscode/mcp.json in your workspace. Start the server from the MCP panel and the SiteLift tools appear in agent mode.

{
  "servers": {
    "sitelift": {
      "type": "http",
      "url": "https://api.sitelift.io/mcp",
      "headers": { "Authorization": "Bearer slk_your_api_key_here" }
    }
  }
}

Any other client

Windsurf, Zed, Hermes and custom agents all work the same way: add a remote MCP server at https://api.sitelift.io/mcp with transport Streamable HTTP and the header Authorization: Bearer slk_your_api_key_here. Clients that support OAuth discovery can sign in instead of sending a key.

Tool reference

You never call these by name; the assistant picks them from your request. The list is here so you know what is on the table.

Limits and the free tools server

Each key or signed-in connection may make 60 requests per minute. Over that you get a 429 with a Retry-After header, which every listed client honours on its own. Long-running work (audits, AI visibility runs, competitor page refreshes) is queued and read back with the matching get_* tool, so no call blocks on a worker.

The free SEO tools are also available over MCP without an account at https://api.sitelift.io/mcp/free (30 requests per minute per IP). Point any client at that URL with no header to run AI readiness, sitemap, robots and the other checks from chat.