I kept rebuilding the same resume parser… so I turned it into an API (looking for testers)

Every time I started a project involving resumes an ATS, candidate matcher, or recruitment automation the first 1-2 days were always the same:

Then repeat it all again for the next project

After doing this a few time, I realized the problem isn't "parsing resumes", it's how inconsistent and time consuming the pipeline is around it.

So i built an API to handle the entire flow

What it does:

Upload a resume (PDF or DOCX)
Runs extraction + structuring using Claude
Returns consistent JSON:
name, email, phone, skills, work experience, education
Usually in ~3–5 seconds

No prompt tuning, no cleanup layer, no schema mapping on your side.

Why not just use AI directly?

You can and I did.

But the issue wasn’t calling an LLM. It was everything around it:

This just removes that layer completely.

What’s live right now:

Upload a PDF or DOCX resume, get back structured JSON in under 5 seconds
Extracts name, email, phone, skills, work experience, and education consistently
Handles messy formatting, missing fields, and inconsistent layouts

What I don’t know yet:

I’ve tested this on my own datasets, but not at scale with real-world messy resumes.

I’m sure there are edge cases I’m missing and that’s exactly what I want to find.

I'm especially curious whether it handles scanned PDFs, non-English resumes, and heavily formatted templates I suspect those are where it breaks.

What I’m looking for:

I’m looking for ~10 developers building:

If you’re dealing with resumes or messy candidate data, I’d love for you to try it and tell me where it breaks.

I’ll personally help you set it up and debug anything.