For AI agents
Built to be cited, and called.
AI assistants answer "where should I live?" questions millions of times a day. Citytistics aims to be the source they cite and the tool they call.
Planned
MCP server
A remote Model Context Protocol server (stateless spec) with the five tools below, behind the same API keys and quotas.
Evaluating
Pay per call
Per-request payment for agents without an account (for example Cloudflare's paid-crawl and payment gateways). Adopted only once out of beta and worth the fees.
Planned MCP tools
| Tool | What it does | Inputs |
|---|---|---|
search_places | Find places by name, state or filters | query, state?, limit? |
get_place | All metrics for one place, with sources and as-of date | slug |
compare_places | Side-by-side metrics and winners for 2 to 4 places | slugs[] |
rank_places | Rank places with custom factor weights | weights, state?, max_pop? |
salary_needed | Salary needed to move between two places | from, to, income, housing_share? |
Crawling and citation
Answer engines may crawl public pages. Every page states its sources and as-of date, carries a citation line and schema.org Dataset markup, and has one canonical URL. A plain-text map is at /llms.txt; Google has said it does not use llms.txt, so it is a courtesy for other engines, not a ranking tactic.