Citehound

cited-geo-guide

This repository holds Citehound, an end-to-end service for AI readiness: a free scanner, Pro site reports, Fix (we make the technical changes) and Beyond (we write the content). The repository and the Vercel project keep their original names; see docs/rename-notes.md.

Benchmark data

The homepage benchmark section (and the "average of N sites" line on every scan report) is generated, not hand-written. To add a category:

1. Scan it: node scripts/scan-category.js domains/<category>.txt <category> — writes data/<category>-raw.json (per-domain results — gitignored, never published) and data/<category>-summary.json (aggregate stats only, no company names).
2. Add one entry to data/benchmarks.json: category (matches the summary filename), label (display name), track ("B2B SaaS" or "Consumer & e-commerce"), note (the unit word for the sample size, e.g. "platforms").
3. Run node scripts/generate-benchmarks.js — rewrites the block in index.html between the BENCHMARKS:START/BENCHMARKS:END markers. Safe to re-run; it replaces the block in place.
4. Commit data/benchmarks.json, the new data/<category>-summary.json, and the regenerated index.html. Leave -raw.json uncommitted — it's gitignored on purpose.

No score or count is ever typed by hand into benchmarks.json or index.html — both are read from the summary file at generation time, so they can't drift apart. Two minutes, no HTML editing.