Send your PDFs
Upload files, forward them to a dedicated email address, or push them through the API. Digital exports, flatbed scans, and photos of handwritten pages all work the same way.
system readyPDF in, structured data outAny PDF in.
Any PDF in.
Clean data out.
Automatically.
Drop in any PDF — digital, scanned, or handwritten — and datahone reads it like a person, lifting fields, tables, and line items into clean, structured JSON. No template to start, handwriting read first-class at no upcharge. Delivered by webhook, the REST API, or a JSON, CSV, or XLSX download.
No card · 20 free pages / moDigital, scanned & handwrittenTables as arrays
stop copy-pasting from PDFs
What is PDF parsing?
A PDF is built to be read by people, not machines — which is why pulling data out of one usually means copy-paste. PDF parsing turns that document into structured, editable data. datahone goes past classic OCR: instead of just reading characters, it understands what each part of the page means.
Digital export, flatbed scan, or a handwritten note — it reads them all, lifts the fields and tables you care about, and hands them back as clean JSON your tools can ingest. Handwriting is first-class, at no upcharge, and you don’t build a template to begin.
Any PDFDigital, scanned, or handwritten — one model reads them all, with no per-document template to set up.
Tables as arraysMulti-row tables and line items come back as structured arrays — however many rows the page runs to.
Checked, not guessedEvery field carries a confidence score; low-confidence ones wait in a review queue before anything posts.
simple workflow
How PDF parsing works. Three steps.
Send datahone a PDF, tell it where the data should go once, and every document after that runs on autopilot. You only ever touch three surfaces.
AI reads the page
datahone handles varying layouts, scanned and handwritten pages, and complex tables — pulling the exact fields you describe, or inferring the shape. Low-confidence fields wait for review.
doc_typePurchase order
referencePO-44120
line_items7
confidence0.98
Export anywhere
Clean JSON streams to any webhook you run — in real time. Pull it from the REST API, download CSV or XLSX, or review it in the dashboard before it lands.
GETapi/v1/exports?format=csv202 queued
GETapi/v1/documents/:id/output200 ok
POST/datahone/webhook200 ok
flexible data capture
Watch it read a PDF.
Point datahone at any PDF and it lifts every field into structured JSON — header data, tables, line items, even a handwritten note. Describe the fields you want, or let it infer the shape.
PURCHASE ORDERHarvest Supply Co.
42 Mill Lane, Bristol, UK
42 Mill Lane, Bristol, UK
Deliver to
Northgate Warehouse
8 Dock Road, Cardiff, UK
| Item | Qty | Unit | Total |
|---|---|---|---|
| Oak board 18mm | 40 | 12.00 | 480.00 |
| Pine batten 2m | 120 | 1.80 | 216.00 |
| Wood glue 5L | 6 | 14.00 | 84.00 |
Subtotal780.00
Order total£936.00
approvedJ. Okafor — ship Fri
document.jsonextracting
{ "docType": "purchase_order", // 0.99 "reference": "PO-44120", "supplier": "Harvest Supply Co.", "orderDate": "2026-04-18", "deliverTo": "Northgate Warehouse", "lineItems": [ { "item": "Oak board 18mm", "qty": 40, "total": 480.00 }, { "item": "Pine batten 2m", "qty": 120, "total": 216.00 }, { "item": "Wood glue 5L", "qty": 6, "total": 84.00 } ], "orderTotal": 936.00, "currency": "GBP", "approvedBy": "J. Okafor", // handwritten · 0.93 "note": "ship Fri", "status": "review_passed"}
what datahone pulls from a PDF
Any page. Any condition. Same clean data.
From a crisp digital export to a coffee-stained scan with a note in the margin, datahone returns the same structured record your tools can use.
Digital & scanned
Native PDFs, flatbed scans, and phone photos are read the same way — deskewed, denoised, and understood, even when the source is far from pristine.
Handwriting, first-class
Handwritten forms, margin notes, and signatures are read in the same pass — at no upcharge. Low-confidence reads route to review rather than guessing.
Tables as arrays
Multi-row tables and line items come back as structured arrays — however many rows the page runs to, and even when columns shift between documents.
Any document type
Purchase orders, delivery notes, statements, contracts, forms, reports, bills of lading — describe the fields you want, or let the AI infer the shape.
Multi-page & bulk
Long, multi-page PDFs and whole batches process in parallel. datahone splits, reads, and stitches the result back into one clean record per document.
Normalised output
Dates, currencies, and number formats come back in one consistent shape — ISO dates, decimal amounts, currency codes — ready to drop straight into your tools.
let AI do the boring work
Why teams run PDFs through datahone.
It puts document data capture on autopilot, so your team spends its time on the work that actually needs judgement.
Hours back, every week
Stop copy-pasting from PDFs. datahone reads a full document in seconds and posts it for review, so the stack that used to take a morning clears itself.
Seconds per document, not minutes
Fewer keying errors
Every field carries a confidence score; anything uncertain is held in a review queue for a person to confirm. Nothing questionable reaches your systems unchecked.
Review queue on low confidence
Lower cost to process
Handle peak-period volume without temp staff or overtime. You pay per page with a hard cap on every tier — overages are blocked, never billed by surprise.
Per page, hard-capped
Secure & compliant
Documents and the data inside them are encrypted in transit and at rest, scoped to your account, and never used to train shared models. EU-hosted, GDPR-aligned.
EU-hosted, encrypted
going further
Pointed at a specific document?
The same engine, tuned for the documents teams parse most. Start on a focused page if you know your doc type.
Solution
Invoice OCR
Tuned for invoices — supplier, line items, totals, VAT, dates, and PO numbers, straight to your accounting tool.
Explore SolutionResume parser
Turn CVs into ATS-ready candidate records — name, contact, skills, work history, and education — in bulk, any layout.
Explore SolutionEmail parser
Turn any inbox — orders, leads, receipts, bookings — into structured data, delivered by webhook or the REST API.
Explore PlatformDocument parsing
The umbrella capability behind every solution — any document in, clean structured data out. See how datahone reads and structures your files.
Explorequestions
PDF parsing, answered.
Q · 01
Digital PDFs, flatbed scans, and phone photos — purchase orders, delivery notes, statements, contracts, forms, reports, bills of lading, and more. Whether it's a clean export or a creased scan, datahone reads it the same way.
Q · 02
Yes — handwritten forms, margin notes, and signatures are read in the same pass as printed text, and at no upcharge. Where a scrawl is genuinely ambiguous, that field is flagged low-confidence and held for review rather than guessed.
Q · 03
Three ways: upload files in the dashboard, forward documents to a dedicated email address, or push them through the REST API. Long, multi-page PDFs and whole batches are all fine.
Q · 04
No. There's no template or rule to set up before you start — describe the fields you want, or let the AI infer the shape. One model reads each layout in context, so a new format doesn't break anything.
Q · 05
Yes. Multi-row tables come back as structured arrays — one object per row, with each column as a field — however many rows the page has, and even when the column layout shifts between documents.
Q · 06
To any webhook endpoint you run, pulled from the REST API, or downloaded as JSON, CSV, or XLSX, ready for the database, app, or warehouse you already use. Review it in the dashboard before it lands. Most teams wire up a destination in minutes.
Q · 07
Every field comes back with a confidence score; low-confidence fields are held in a review queue for a person to confirm before anything posts. Dates, currencies, and numbers are normalised to one consistent shape, in any language.
Q · 08
Per page, with a hard cap on every tier — one PDF page is one page. When you hit your limit the next page is blocked rather than silently billed, so a bulk run can never produce a surprise bill. See pricing →
get started
Stop copy-pasting from PDFs.
Drop in any PDF — digital, scanned, or handwritten — and watch datahone hand it back as clean, structured data, routed wherever you need it. Free in minutes, no card required.
Digital, scanned & handwrittenTables as arraysRouted anywhere