# Clickport

Clickport is privacy-first, cookie-free web analytics. It tracks visitors without cookies or personal data, and its API also reports which AI agents read a site: what they crawled, what they ingested, and what they cited.

## Documentation

- [Clickport docs](https://clickport.io/docs): Start here for the dashboard, tracking setup, and every feature.
- [API reference](https://clickport.io/docs/api): Read analytics, ingest events and revenue, and pull AI agent data over HTTP.
- [Show Clickport data inside your product](https://clickport.io/docs/show-data-in-your-product): For agencies and platforms with many client sites: create a site per client with one team API key, install tracking and the AI agent connector, read each client's numbers, and handle request and site limits.
- [AI agent data](https://clickport.io/docs/ai-agents): How Clickport detects and classifies AI agents, and what each site can see about them.
- [MCP server](https://clickport.io/docs/mcp): Connect an LLM client to a Clickport site over the Model Context Protocol.
- [OpenAPI document](https://clickport.io/api/openapi.json): Machine-readable description of every API route, its parameters, and its required scope.

## Vocabulary

- reads / crawled: any verified fetch of a page by an AI agent.
- ingested: the agent's intent was indexing or training. The page was added to a search index or used to train a model.
- cited: the agent's intent was live_retrieval. It fetched the page during a live user conversation. This is the closest observable proxy for a citation. Clickport sees the fetch, not the answer the agent gave.
- visits / conversions: human visitors who arrived from an AI referrer, and what they did after. This is the demand side of AI traffic, separate from the fetch side above.
- spoofed: a fetch that claimed an AI agent's user agent from an IP address outside that agent's published ranges. Excluded from every number above unless a caller asks for it.