Stract API: Turn Documents into Structured Data

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!