MCP for AI agents

Connect Claude, Claude Code, Cursor and other MCP clients to the hosted AdLib MCP server at mcp.adlibapi.com and let your agent research ads.

The AdLib MCP server gives an AI agent 20 tools over the ad libraries: resolve a domain, search ads, read history, check whether a brand advertises, and set up watchlists. It is hosted, so there is nothing to install.

Output

https://mcp.adlibapi.com/mcp

Claude.ai and Claude Desktop

Open Settings → Connectors → Add custom connector, name it AdLib, paste the URL, click Connect and sign in with your AdLib account.

Claude Code

Shell

claude mcp add --transport http adlib https://mcp.adlibapi.com/mcp

Then run /mcp in Claude Code and choose adlib to sign in. With an API key instead:

Shell

claude mcp add --transport http adlib https://mcp.adlibapi.com/mcp --header "X-API-Key: your-key"

Cursor and other clients

Add the URL as a remote MCP server (in Cursor: {"mcpServers": {"adlib": {"url": "https://mcp.adlibapi.com/mcp"}}}) and sign in, or send your key as the X-API-Key header.

Prompting

Name AdLib in the prompt, so the agent uses these tools rather than a web search:

"With AdLib: which platforms is hellofresh.com advertising on right now, and since when?"
"Use AdLib to compare the ad copy Nike and Adidas ran on Meta in September."

Every tool result ends with the credits it used, for example billed: 5 credits. Empty results and errors are free, as in the REST API.