A Mahad technology product
Read any document. Get clean data back.
Passports, national IDs, visas, residence permits, CVs, invoices, receipts and certificates — read into structured JSON through one API. Every field is marked ✓ confirmed or ⚠ check, with the evidence behind it.
- Never invents a value
- English and Arabic first
- Free plan to start
Example
How a result looks
Pick a sample. Each field shows ✓ with its evidence, or ⚠ with the reason a person should check it.
Result
- Surname Confirmed
AL-MARRI
Proven by the MRZ check digits
- Given names Confirmed
TARIQ KHALID
Printed page and MRZ agree
- Passport number Confirmed
N08492019
Proven by the MRZ check digits
- Date of birth Confirmed
1991-11-14
Proven by the MRZ check digits
- Date of expiry Confirmed
2031-08-12
Proven by the MRZ check digits
- Place of birth Check
DOHA
Two readers disagree — please check
Illustration with fictional sample documents — try the live demo with your own file.
Documents
What Mahad OCR reads
Any document, any country. Each type has its own page with the fields returned and its current limits.
Passports
Printed page and machine-readable zone. MRZ values are proven by the MRZ check digits.
National IDs and residence permits
Front and back of ID cards and residence permits from many countries.
Visas
Visa pages and entry permits: names, numbers, dates and issuing country.
CVs
Old, scanned or photographed CVs turned into a clean structured profile.
Invoices and receipts
Supplier, numbers, dates, totals and line items, with totals checked against the lines.
Certificates and other documents
Diplomas, letters and general documents read into text and fields.
How it works
Upload, read, review
From a file to structured data your systems can use.
1. Upload
Send a photo or file (PNG, JPG, PDF, up to 10 MB) from the web app or the API.
2. Read
Mahad OCR reads every field and returns clean JSON. Each field carries the evidence behind it.
3. Review
Fields marked ⚠ go to a built-in human review step. Your team confirms or corrects them.
Why Mahad OCR
Clear about what it knows
Never invents a value
Unclear fields are marked ⚠ with a reason and left for a person to check — not guessed.
Evidence on every field
Each ✓ says why: two readers agree, the MRZ check digits prove it, or it was taken from the file text.
English and Arabic first
Names in both scripts are read side by side. Other scripts vary — test with your own documents.
One JSON API
One key, one header (X-MahadDoc-API-Key), the same result shape for every document type.
CVs
Old, messy CV → clean profile
Phone photos of paper CVs, screenshots and old Word files become one structured profile.
- Reads contact details, experience and education.
- Missing details stay empty — never invented.
- Returns JSON for your system, or a clean CV file.
Developers
One JSON API
Create an account, generate a key (it starts with mk_live_ and is shown once), upload a file and read the result. There is a free plan to start.
{
"document_type": "passport",
"values": {
"surname": "ERIKSSON",
"document_number": "L898902C3",
"date_of_birth": "1974-08-12"
},
"field_evidence": {
"surname": "corroborated",
"document_number": "deterministic",
"date_of_birth": "deterministic"
},
"review_reasons": []
}Built on privacy and honesty
Your documents are kept to your own account, API keys are stored hashed, and every important action is written to an audit log. Where the software is unsure, it says so with ⚠ instead of guessing.
confirmed check itMahad OCR is built and operated by Mahad Information Technology.