Built by FlintHorn

The First Judicial Circuit of Florida came to us with thousands of administrative orders sitting as PDFs. They needed them published as accessible web content. The right solution wasn't a custom script — it was a reusable pipeline.

We built Pith to solve that problem: transform unstructured documents into structured CMS content using AI, with a review step that makes sure every field is right before it ships.

Every feature in Pith comes from a real workflow problem. Confidence scores exist because reviewers need to know where to focus. Reference resolution exists because manually cross-linking related documents at scale is impractical. Batch processing exists because court systems don't have five-document backlogs.

We're a small team and we build deliberately. If you're working on a document processing problem we haven't seen, we want to hear about it.