One search for car fines across 33 government systems that don't talk to each other
Hey PH 👋 We launch Multita in a few hours and I wanted to open a thread for anyone curious about the build, or anyone who's fought the same fight with fragmented public data.
The problem: in Argentina, the traffic fines for a single car can live across 33+ separate government systems. One per province, plus dozens of municipalities. None of them share data, each has its own portal, its own captcha and its own random downtime. Checking one car by hand means opening four or five of them and burning 15 to 20 minutes per vehicle.
Multita (https://multita.com.ar) turns that into one query. You enter a license plate, a national ID or a tax ID, and it returns every fine across every jurisdiction, with amount and status, in seconds. It's free for anyone. For the paperwork agencies (here we call them "gestorÃas") that do this all day, it also quotes the debt with their own branding and exposes a JSON API.
A few things we learned the hard way:
No two portals agree on anything. Date formats, what "paid" means, even what counts as a fine. Normalizing it all into one clean response was most of the work.
Captchas are the real boss fight. One provincial portal runs reCAPTCHA v3 that rejects every non-human token, so we just don't claim coverage we can't deliver. Honest gaps beat fake data.
Reliability beats breadth. A source that returns wrong numbers is worse than one we skip, so we gate anything we can't fully trust.
Two questions for the PH crowd:
If you've built on top of messy public data (fines, taxes, court or utility records), how did you tame the "every source is its own snowflake" problem?
For the API people: would you rather receive raw normalized JSON, or totals with the business logic already baked in?
Happy to get into the stack, the scraping, or the business side. 🙌
Replies