What it is, and why.
SaúdeCidade maps Brazilian healthcare: clinics, hospitals, dentists, labs and public health units — 556,745 establishments across 500 cities, sourced from the official CNES/DataSUS registry rather than scraped listings.
The first production version went from zero to deploy in 7 days, reusing the directory engine I had already proven on AutoCidade: same Go binary, same LevelDB layout, different ingestion and editorial layer.
Designed, built and operated end to end by me — from ingestion to the page you see at saudecidade.com.
3 decisions that shaped it.
Official data, not scraped data
Ingestion runs against the CNES/DataSUS API, so every establishment carries a government registry identity. That makes the directory verifiable and keeps it correctable when the source updates.
Editorial layer people actually search for
Around 100 guides on vaccination schedules, consultations, specialties, exams, SUS access, emergency care, dentistry, physiotherapy, nutrition and mental health — organized into topic clusters with pricing tables and FAQ schema.
Health-plan analysis with real numbers
ANS rating data is parsed and de-duplicated to answer questions the raw ranking cannot: 20 plans tie at the maximum score, 244 distinct Unimed entities compete under one brand, and only about 8 genuine quality tiers exist underneath.
Same engine, other cities and markets.
Want something like this built?
Directories, content engines and the data behind them. Tell me what you need to prove, and to whom.