* * * PRICE CHECK * * *
PennyOCR vs Google Cloud Vision.
Cloud Vision's DOCUMENT_TEXT_DETECTION is solid OCR wrapped in a big-cloud console. Here's how it stacks up against PennyOCR — list prices as of August 2026.
| PENNYOCR | GOOGLE CLOUD VISION | |
|---|---|---|
| PRICE / 1,000 PAGES | $0.75, every tier | $1.50 (document text detection, units 1,001–5,000,000) |
| TABLES | Included — HTML tables inside markdown | No native table structure; you rebuild from word boxes |
| FREE TIER | 100 pages / month, forever | First 1,000 units / month |
| SETUP | Email code → API key → curl. ~2 minutes | GCP project, billing account, service-account JSON |
| OUTPUT | Markdown or plain text, reading order preserved | JSON with per-word bounding boxes and confidence |
| BILLING | Prepaid credits, never expire | Postpaid on your GCP bill |
| DATA RETENTION | Zero retention by default | Per Google Cloud data-use terms |
Competitor figures are public list prices and terms as of August 2026 — check their pricing pages for changes.
CHOOSE PENNYOCR IF…
Choose PennyOCR when the goal is clean text/markdown from documents — half the price, tables included, and no service-account ceremony.
CHOOSE GOOGLE VISION IF…
Choose Cloud Vision when you're GCP-native or also need its other vision features (labels, faces, landmarks) in the same call.
FAIR QUESTIONS
DOES PENNYOCR RETURN BOUNDING BOXES?
Not today — we return clean text/markdown in reading order, which is what most pipelines actually consume. If you need per-word geometry, Vision fits better.
HOW DO THE FREE TIERS COMPARE?
Vision gives 1,000 units/month; PennyOCR gives 100 pages/month plus prepaid credits that never expire, with no card required to start.
WHICH IS BETTER FOR RECEIPTS?
We're literally named after them. Our VLM handles crumple, fade and skew well — and our receipt benchmark and eval harness are public, so you can verify rather than trust.