Launching today
Stract API
Turn Documents into Structured Data
2 followers
Turn Documents into Structured Data
2 followers
Upload PDFs, office docs, or scans, describe the data you need in plain language, and get CSV or JSON back. No manual entry, no brittle templates.
Hey Product Hunt, I'm Matt, the maker of Stract API. 👋
Why I built it: I kept running into the same problem. At a construction tech startup, we used an LLM to turn a crew's plain-language description of their day into structured labor and material entries. It worked, but it was fragile: parsing the output was painful, bad output was constant, and any change to the structure broke it. Later, at a venture automating land title research, I spent weeks building a flexible way to pull data out of dense legal documents. Both times, extraction became a core part of the product and pulled me away from what we were actually trying to build. I just wanted a service to do it for me. When I asked a Google exec at a startup accelerator whether one existed, he paused and said no, but that they probably should build one. So I did.
Stract API works like this: you describe the fields you want (strings, numbers, enums, dates, and so on), send a document, and get back CSV or JSON that matches your spec. It has a point-and-click web app and a REST API with synchronous endpoints. You can try it with 100 free credits, no card required.
Feedback that would help me most:
1. What documents are you trying to extract from? I want to know which types or domains we're not handling well yet.
2. Long documents. It does well up to about 20k words (~40 single-spaced pages). I'm building an agentic approach for longer ones, so I'd love to hear where you hit limits.
3. Speed. Jobs take about 5 to 30 seconds per document today. Is that fast enough for your workflow, or do you need it faster?
4. SDKs. If you'd use one, which languages?
I read every reply. Thanks for checking it out!