A team working county court records in the US. October to November 2023.
County court records, retrieved automatically
Teams that depend on public records pay people to run the same searches every day and download documents one at a time. We drove the Hamilton County foreclosure search through Selenium, retrieved every matching filing with retries, and returned the result in a shape the team could work with.
The county foreclosure search, run and downloaded by a script
Screens from this system are being added.
The situation
Teams that depend on public records pay people to run the same searches every day and download documents one at a time. The team needed the Hamilton County foreclosure search run repeatably, every matching filing retrieved, and the result in a shape it could work with.
What we built
Automated search against the county clerk-of-court system, driven through Selenium
Matching PDF retrieval and structured output
Retries on PDF retrieval, so a failed download is re-attempted within the run
Repeatable runs, so the same search is not re-done by hand
What changed
The same search stopped being re-done by hand, and a failed download was retried within the run instead of noticed later.
An email becomes order lines against a catalogue, stock moves to reserved, and a quotation drafts itself. 3 quantities per stock item: on hand, available and reserved.
The CSV or JSON export an ads-ops person opens every morning, aggregated per campaign, compared against the previous day, with what to look at first pulled to the top. 40 minutes of a person's morning, and a spreadsheet, before the prototype.
Gemini 2.5 Pro reads a scanned Thai and English insurance form and returns every field as structured JSON, on video, with the SAP entry stage mapped for the operation to decide on. 46 fields out of one scanned insurance form, as JSON.