Skip to main content

​
What Is the LanderLab MCP Server?

The LanderLab MCP server connects your AI assistant to your LanderLab account using the Model Context Protocol. Once connected, your assistant can list landers, pull analytics, publish pages, manage leads, run A/B tests, and more without you ever opening the dashboard. MCP server URL:
  • Sign in with your LanderLab account (OAuth)
  • No API key, no headers, no config file editing
  • 30+ available tools
  • Works with Claude, Cursor, Windsurf, Manus, ChatGPT, and any MCP-compatible client

​
Who Is This For?

Performance marketers, media buyers, and teams who want to move faster. Instead of switching between your AI tool and the LanderLab dashboard, you describe what you need in plain English and the assistant handles it.

​
How It Works

The MCP server exposes tools that your AI assistant calls during a conversation. When you say “list my landers in workspace 1,” the assistant calls landers_list and returns the results directly in your chat.

​
Connect Your AI Assistant

Choose your tool and follow the setup guide: Claude finds LanderLab in its connector directory. Any other MCP-compatible client can connect by pointing to https://api.landerlab.dev/mcp and completing the sign-in prompt.
What about Grok, Perplexity, and Copilot? These tools do not currently support connecting to external MCP servers as a client. Grok and Perplexity can act as MCP servers, exposing their own search tools to other clients, but their chat interfaces cannot connect to a server like LanderLab MCP. Microsoft Copilot Studio does support MCP but requires a complex enterprise setup that is outside the scope of this guide.

​
Tools Reference

Your organization is resolved automatically from the account you sign in with. You do not need to pass an organization ID.

​
Read-Only Tools

​
Mutation Tools

​
Example Prompts

Once connected, talk to your assistant in plain English: “List all my landers and their status” Calls landers_list and returns page names, URLs, and publish status. “Show me analytics for lander X over the last 7 days” Calls analytics_stats with the date range and returns visits, conversions, and leads. “Unpublish lander X” Calls lander_unpublish to take the page offline. “Set A/B test weights to 70/30 for lander X” Calls variants_set_weights to adjust traffic split between variants. “Show me all leads from workspace 1” Calls leads_list and returns lead data for that workspace.

​
Authentication

The MCP server uses OAuth. The first time your assistant connects, it opens a LanderLab sign-in page in your browser. You approve the access request once, and your client stores a token it refreshes on its own. Nothing sensitive lives in your config files, and you can disconnect at any time from Settings > Connected Apps in LanderLab, or by removing the connector in your AI tool.
API keys are still used for direct REST API calls. See Generate an API Key. They are no longer used for MCP.

​
Troubleshooting

The sign-in window does not open - Some clients block pop-ups on the first connection attempt. Allow pop-ups for your AI tool and try connecting again. Authorization succeeded but no tools appear - Restart your AI tool. Most clients only load the tool list when a session starts. 401 Unauthorized - Your session expired or access was revoked. Remove the connector and add it again to trigger a fresh sign-in. Not Acceptable - The Accept header must include both application/json and text/event-stream. This error usually appears when testing with curl or a custom client. Still pointing at the old URL - If your config still contains https://api.landerlab.dev/api/v2/mcp and an X-API-Key header, delete that entry and reconnect using https://api.landerlab.dev/mcp.