Greek receipt OCR
Receipt OCR that actually reads Greek.
Most capture tools treat Greek as an afterthought, mangled characters, missed ΦΠΑ lines, supplier names half-transliterated. Pileform was built in Cyprus with Greek as a first-class language: it reads the receipt the way your reviewer does.
The Greek receipt problem
A Cyprus or Greece quarter is full of Greek paper: thermal supermarket tickets, χειρόγραφα taxi slips, supplier invoices where the header is Greek and the line items are English. Generic OCR garbles the Σ and the ώ, skips the «συμπεριλαμβανομένου ΦΠΑ» phrasing that decides whether a total is gross or net, and outputs rows someone has to retype anyway.
11
Languages read natively, Greek included
2
VAT rate systems built in: Cyprus and Greece
~12 min
For a typical 200-page batch
Pileform reads Greek natively, characters, VAT phrasing, and layout, and flags anything it could not read with confidence, instead of guessing quietly.
How it works
Greek in, clean workbook out.
Drop the whole period as one PDF, Greek, English, or both on the same page. The extraction engine reads 11 languages; Greek is not a plugin, it is one of the languages the engine was tuned on from day one.
Drop the period's pile.
Scan or collect the period's receipts and invoices into one PDF, Greek, English, or mixed on the same page, and upload it once.
Greek read natively.
Characters, ΦΠΑ phrasing, and mixed-script lines are read the way a Greek-speaking reviewer would read them, not transliterated or guessed.
Cyprus or Greece rates, per line.
Cyprus 19/9/5/3/0 and Greece 24/13/6 are both in the engine, applied from what each line prints. A missing rate is inferred and yellow-flagged, never silently assumed.
One workbook per supplier.
Source Greek text is preserved for audit; workbook headers come out in your firm's configured language.
ΦΠΑ phrasing, understood.
«Συμπεριλαμβανομένου ΦΠΑ», «πλέον ΦΠΑ», «ΦΠΑ 19%», the phrases that decide whether a total already contains the VAT. Pileform tags every total as inclusive or exclusive based on what the document actually says, and preserves the source phrase in the workbook for audit.
Σύνολο €119,00, συμπερ. ΦΠΑ 19% books as gross 119.00, net 100.00, VAT 19.00, not as net 119 with VAT added on top.
Cyprus and Greece rates, native.
Cyprus 19 / 9 / 5 / 3 / 0 and Greece 24 / 13 / 6 are both in the engine, applied per line from what the receipt prints. A rate printed on the document is never overridden; a missing rate is inferred and yellow-flagged for review.
Greek and English on the same receipt.
A Cyprus supplier invoice with a Greek header, English line items, and a bilingual VAT summary is normal paper here. The extractor handles mixed-script documents line by line, no language toggle, no separate batches per language.
- EN · EL · DE · FR · ES · IT
- NL · PL · RO · BG · RU
The sample workbook on the features page includes Greek-language receipts, see the rate column, the inclusive/exclusive tags, and the embedded source images before you sign up.
How it works
The whole period, one drop.
The whole period, one drop.
Not photo-by-photo capture. Scan the pile into one PDF, hundreds of pages, mixed suppliers and languages, and drop it once. A typical 200-page batch processes in about 12 minutes; you get an email when the workbook is ready.
Greek sources, your headers.
Workbook headers come out in your firm's configured language; the source words stay in Greek where the document was Greek, preserved in cell comments for audit. One workbook per supplier, with the original receipt image embedded next to its rows.
ΦΠΑ phrasing, understood.
«Συμπεριλαμβανομένου ΦΠΑ», «πλέον ΦΠΑ», «ΦΠΑ 19%», the phrases that decide whether a total already contains the VAT. Pileform tags every total as inclusive or exclusive based on what the document actually says, and preserves the source phrase in the workbook for audit.
Cyprus and Greece rates, native.
Cyprus 19 / 9 / 5 / 3 / 0 and Greece 24 / 13 / 6 are both in the engine, applied per line from what the receipt prints. A rate printed on the document is never overridden; a missing rate is inferred and yellow-flagged for review.
Greek receipts vs. generic OCR
What a plain OCR box misses.
| Pileform | Generic OCR | |
|---|---|---|
| Greek characters (Σ, ώ, ΦΠΑ) | Read natively, no transliteration | Garbled or dropped |
| «Συμπεριλαμβανομένου ΦΠΑ» phrasing | Tagged inclusive/exclusive from source | Not detected |
| Cyprus 19/9/5/3/0 + Greece 24/13/6 | Applied per line, printed rate wins | No rate logic |
| Greek and English, same receipt | Read line by line, no toggle | One language per batch |
| Handwritten Greek taxi slip | Retried, yellow-flagged if unsure | No signal |
See it on a real Greek receipt.
The sample workbook on the features page includes Greek-language receipts, see the rate column, the inclusive/exclusive tags, and the embedded source images before you sign up.
Greek receipt OCR questions, answered.
Partially, honestly. Printed and thermal-printed documents are the focus, and that is where the extraction is strong. Handwritten items, the classic Greek taxi slip, are attempted and retried automatically, but anything the engine is not confident about comes back yellow-flagged for human review rather than silently guessed. You will know exactly which rows to check.
Handled natively. Mixed-script documents, Greek header, English line items, bilingual VAT summaries, are read line by line without any language setting. The engine reads 11 languages and detects the language per line, not per document.
Yes. Greece’s VAT rates (24% standard, 13% and 6% reduced) are supported the same way as Cyprus’s 19 / 9 / 5 / 3 / 0, applied per line from the printed rate, with inference flagged. Greek-language documents behave identically regardless of which side of the Aegean they came from.
Yes. Bank statements go into the same period drop as the receipts and invoices, and transactions come out as double-entry rows reconciled within the same workbook. Greek-language statement layouts from Cyprus banks are part of normal processing.
Scan or collect the period’s documents into one PDF, order and orientation do not need to be perfect, and upload it. Pileform processes pages in parallel, retries the hard ones, and emails you when the workbook is ready. A typical 200-page batch takes about 12 minutes; larger batches finish overnight.
Drop your messiest Greek-language pile.
Sign up free, no card. The first batch is on us, see how the Greek comes out before you commit to anything.