Free 40-page Claude guide — download today
Paid · instant download

Receipt / Invoice Data Extractor — image → structured JSON

Upload an image of a receipt or invoice, get structured JSON with line items, tax, totals, and vendor info. Free DONUT or LLaVA + Tesseract path.

Commercial path: Google Document AI ($1.50 per 1000 pages). This wrapper ships with the free open-source alternative: DONUT or LLaVA + Tesseract (free, local GPU helpful).

Stack: Python / FastAPIOSS difficulty: advancedLifetime updates
Walkthrough video

Live demo would need a GPU.

This wrapper runs an image-generation or vision model that needs GPU compute the demo VPS does not have. Instead, the wrapper ships with a Loom walkthrough showing the full setup and output. You run it on your own machine (or a $0.30/hr GPU rental) — full setup in the included guide.

Loom walkthrough link ships in the download. Sample outputs are included in the ZIP under sample-outputs/.

What's in the ZIP

Code + guide + samples. Everything to ship in 30 minutes.

  • Python FastAPI app
  • Structured JSON with line items + tax + totals
  • DONUT + LLaVA + Tesseract pipeline
  • Swap-in guide: Google Document AI → local
  • Layman-language guide on the OSS swap (no jargon, no assumed expertise)
  • README with copy-paste setup commands
  • Sample inputs and reference outputs
Why include the OSS path

Because the wrapper is the part you keep.

Most AI starter kits hardcode Google and call it a day. That puts you on a recurring bill for every demo, every test, every customer. The wrapper here ships with the working Google path AND the free open-source path side-by-side. Use the paid API in production if you want the convenience. Switch to DONUT for development, testing, or cost-sensitive deployments. Same code, different endpoint.

The guide walks through the swap in layman language — no ML jargon, no assumed familiarity with vector databases or quantization. If you can run npm install or pip install, you can follow it.

Common questions

When do I actually get the code?
Instantly after payment. PayPal / Razorpay confirms, you get an email with a private unlock link, the ZIP downloads in one click.
Do I need to know AI/ML to use this?
No. The wrapper is a working app you can deploy in 30 minutes following the README. The OSS-swap guide is written for developers who know JavaScript or Python but have never touched a model file. No PhD needed.
What does "free open-source model" actually mean here?
DONUT or LLaVA + Tesseract (free, local GPU helpful)is a model (or set of models) you can run on your own hardware with no API keys, no per-call billing, and no rate limits beyond what your CPU/GPU allows. The guide explains how to install, configure, and swap it into the wrapper's existing code path.
How does this compare to building it myself?
You could absolutely build this yourself. The wrapper saves you 8-20 hours of plumbing — API client setup, prompt engineering, error handling, the OSS swap research, sample data, testing. If your time is worth more than $1/hour, the math works.
Lifetime updates?
Yes. When v1.1 ships with new features or OSS-model paths, the same unlock link serves the latest version.
Refund policy?
Digital product, all sales final. If something genuinely does not land for you, reply to your purchase email and I will add you to the full Skills Library (lifetime access) as a goodwill gesture.
Buying more than one?

All 15 paid wrappers for $99.

Saves $201 vs buying individually. Every future wrapper added at no extra cost.

See the bundle →