DocuneX reads scans, PDFs, photos and Office files, keeps every heading, table and line in place, then lets you ask questions with page citations, export clean data, and find the best prices for what is on a bill or prescription.
Pilot pricing on request. Built for families, small businesses, pharmacies and clinics.
Drop in a scan, a PDF, a photo or an Office file. DocuneX reads it with open-source OCR, rebuilds its headings, lists and tables, and keeps every line where it was on the page. Then ask it a question and get an answer with its page, export clean data, or turn the items on a bill or prescription into a cart at the best price.
Design targets and capacity, measured on demo data — not results from client deployments.
Paper rarely arrives as a clean PDF. DocuneX takes what you have, up to 50 MB, and works out what it is from the file itself. It reads:
Ordinary OCR scrambles columns and flattens tables. DocuneX rebuilds the page the way it was written, and shows it seven ways. It keeps:
Ask the way you would ask a person, and get an answer built only from the document, with the page it came from. Answers come:
Photograph a bill or a prescription and DocuneX lists what is on it, ready to buy, with the best place to buy it. For each item it finds:
Once a document is read, the data is yours to take anywhere. Every view can be copied, and the result exports as:
The paper is usually in your hand, not on your computer. The DocuneX app scans it where it is. On the phone you can:
Drop in a file up to 50 MB. Its type is detected from its contents, so a renamed file is still read correctly.
Digital pages are read from their own text layer; only scanned pages go through OCR, so the slow step runs only where it has to. Five steps take a file to something you can act on.
The reading is done by free, open-source engines on an ordinary CPU; the language model is used only where it adds something a rule cannot.
Next.js and React, with pdf.js for digital PDFs and parsers for Word, Excel and PowerPoint.
Read a handwritten prescription, understand the dosage, and find the medicines at the best price.
Turn photos of bills and invoices into clean rows, without typing them out.
Get to the renewal date, the notice period or the summary of a 34-page report in seconds.
Read prescriptions into medicines, dosages and generic names, ready for the counter.
Turn shopping lists and bills into products matched with stores and prices.
The same pipeline inside your own app or workflow, in your brand.
Documents are often personal: prescriptions, bills, contracts. DocuneX is plain about how each one is handled.
OCR
OCR runs on free, open-source engines, with no paid OCR service in the loop.
Hardware
Measured on a CPU-only laptop: under a second for a digital page, 2 to 25 seconds for a scan.
Storage
There are no accounts and no history: DocuneX does not keep your documents.
AI
Questions and product detection send the extracted text and page images to Google Gemini.
Grounding
Answers come only from the document, each fact with its page, and instructions inside a document are ignored.
Honesty
If the document doesn’t answer the question, DocuneX says so instead of guessing.
One upload, read once, used six ways: from the raw text to an answer, a spreadsheet or a pharmacy cart.
PDFs (digital or scanned), photos, Word, Excel and PowerPoint, CSV and text files, up to 50 MB. The file type is detected from its contents, not its name.
Columns put back in reading order, with headings, paragraphs, lists and tables rebuilt, and every line keeping its position and a confidence score.
Questions answered only from the document, in your own language, with every fact linked to its page. Type the question or say it.
The items on a bill or prescription pulled out by name, brand, quantity and dosage, with spelling corrected and prices compared across Indian stores.
Side-by-side, structure, exact layout, plain text, Markdown and JSON views, with every one ready to copy or export.
A mobile app that scans with the camera or picks a photo or file, and shows the document, its products, answers and text in four tabs.
DocuneX is one of our demo builds. Replace bills and prescriptions with claims, delivery notes, KYC forms or lab reports, and the pipeline carries over.
Scans, phone photos and handwriting read on open-source engines, with a fallback when one fails.
Columns, headings and tables rebuilt from raw text positions, not guessed from the words.
A language model held to the document, with every fact traced to its page.
Free text turned into typed records against a strict schema: names, quantities, dosages, prices.
From reading to a decision: where to buy, what to export, what to do next.
Claims, invoices, KYC, lab reports or delivery notes: the same pipeline applies.
Everything DocuneX does is in six parts. Open any card to see the detail.
READ
Any document
OCR
Two OCR engines
STRC
Structure rebuilt
ASK
Ask the document
BUY
Products and prices
OUT
Export and mobile
DocuneX is a demo build, and we would rather show than claim. We are taking a small number of pharmacies, clinics, retailers and operations teams as pilot partners, to turn it into measured results on their own documents.
One document-heavy process, a written scope and a fixed end date before any work starts.
Reading accuracy, table recovery and answer quality, on your own documents.
Below our standard rates, in return for honest feedback and a decision at the end.
Nothing is published without your written approval: not your name, not your numbers.
Client stories will appear here once pilots finish — real names, with permission.
What people ask first: file types, handwriting, prices, accuracy and where a document goes.
Still have a question?
Book a 30-minute call with a DocuneX engineer. Bring the document your team still types out by hand.
It reads a scan, PDF, photo or Office file, rebuilds its structure (headings, paragraphs, lists and tables), and lets you ask questions about it with page citations, export the result, or find the best prices for the products on it.
Bring one document-heavy process. We’ll tell you straight whether an AI system can read it reliably, and what it would cost to build.
Book a walkthrough
30 minutes on the documents your team types out by hand today.
Bring a sample set
Real bills, prescriptions or reports: we tell you straight how well they read.
Agree the test set
Known values and pages, written down before anything is built.
Pilot in your workflow
A scoped pilot, with a named lead and the results reported honestly.
Looking at another product? See all four
