Introduction

The AIclicks MCP server lets your AI assistant query your AIclicks data directly. No glue code, no custom integration — your assistant gets a typed tool surface over the same /api/v1/* endpoints that power the dashboard.

It speaks the Model Context Protocol, so any MCP-aware client (Claude Desktop, Claude Code, claude.ai web, Cursor, …) can connect to it the same way.

What it gives you

Once connected, your assistant can answer questions like:

• "How visible is my brand in AI search results this week?"

• "Which sources are LLMs citing when they mention us?"

• "Compare our share of voice against competitors over the last 30 days."

• "Which prompts are losing visibility month over month?"

• "Where are we underperforming on Claude vs. ChatGPT?"

…by routing the request through one of the tools described in Tools.

How it works?

• Read-only. No write tools — you can't create domains, add prompts, or trigger analyses from the MCP surface.

• No raw SQL. All queries go through the same /api/v1/* envelope as the public API. Results are cached server-side; you get the dashboard view, not raw rows.

• No background work. Each tool call is a single synchronous request. There are no long-running jobs you need to poll.

How it's hosted

The server is a streamable-HTTP MCP endpoint at:

https://mcp.aiclicks.io/mcp

Authentication uses the same ak_live_* API keys as the public REST API — no separate signup, no separate keys. See Setup for client-specific install steps.

Next steps

• Setup — connect your client (Claude Desktop, Claude Code, claude.ai web, Cursor, others).

• Tools — the full tool reference.

• Recipes — prompts and tool-chain templates you can copy.

• Troubleshooting — common errors and how to fix them.