BENGALI · MARATHI · HINDI

Indian language OCR test documents

Indian language documents. Real layout complexity.

Multilingual extraction needs more than English test files. This free pack contains Bengali, Marathi and Hindi forms with printed fields and handwriting, each paired with a degraded counterpart.

ZIP · 11.9 MB · Free download · No sign-up

10 PDF files5 Document pairsJSON + boxes
INSIDE THE ACTUAL PACKPDF / 01
Bengali savings form · A: readable sample from the Indian Languages packPDF SAMPLE
Same document.
Different condition.

KNOW WHAT YOU’RE DOWNLOADING

Five document pairs.
Ten PDF files.

Explore Bengali, Marathi and Hindi forms with printed text, handwriting and challenging capture conditions. Start with a readable example, then compare its degraded version.

Files in the Indian Languages sample pack
Document typeReadable exampleChallenging version
01Bengali savings form · AClean PDF2 pagesStrong blur2 pages
02Bengali savings form · BClean PDF2 pagesCoffee stain2 pages
03Hindi savings form · AClean PDF2 pagesPhotocopy2 pages
04Hindi savings form · BClean PDF2 pagesLow light2 pages
05Marathi customer formClean PDF2 pagesCropped edges2 pages

10 PDFs plus structured JSON and bounding-box reference files. A file may contain more than one page.

Download the complete pack

FROM DOWNLOAD TO WORKFLOW

Put the sample
to a practical test.

Use a consistent process. Define your expected outputs, run the files and inspect the cases your workflow gets wrong.

01

Multilingual extraction

Check populated fields across Bengali, Marathi and Hindi scripts.

02

Handwriting and form layouts

Test reading order and field localization on multi-page forms.

03

Capture resilience

Compare each source with its degraded counterpart.

RELATED FIELD GUIDEHow to test OCR accuracy on scanned documents

CUSTOM DATASETS · MANAGED SERVICE

Need more languages, layouts or document types? Tell us what your pipeline has to read.

Need custom datasets but don’t want to build and maintain the generation process internally? Let’s discuss your requirements.

Talk to our team

THE NEXT STEP · FREE FOR QUALIFIED TEAMS

Start with 10.
Go further with the full dataset.

Try the sample pack first. When you need more, fill in the booking form and join a call. We’ll review your use case and share the full dataset with qualified teams.

No purchase required. Custom datasets and product support are discussed separately.

We’ll open the calendar for your selected dataset. Add your details in the booking form.

  1. Choose a time and complete the form
  2. Discuss your documents and intended use
  3. Receive the dataset if it fits your team

Booking opens Google Calendar. Privacy policy

BEFORE YOU START

Your questions,
answered.

What will I receive?

A ZIP with 10 PDFs: five clean documents and five challenging versions, plus structured JSON and bounding-box JSON for selected populated fields. Some PDFs contain more than one page.

How do I get the sample?

Choose “Download 10 PDFs”. The ZIP downloads directly: no form, call or platform account is required for the sample.

Can I use my own tools?

Yes. Use the PDFs and reference files in your own evaluation, sandbox or testing workflow. The included reference data is not an automated scoring system.

Are the five forms the same person in three languages?

No. They are five separate fictional applicants across Bengali, Marathi and Hindi, not translations of one record. Each PDF has two pages.

How do I get the full dataset?

Choose your dataset, complete the booking form and join a call. We’ll review your use case and share the full dataset with qualified teams. No purchase is required; availability and fit are confirmed on the call.

FREE FULL DATASET

Let’s find the right data for your team.

Complete the booking form and join a call. We’ll review your use case and share the full dataset with qualified teams.

Your dataset request

Your selection opens its dedicated booking calendar.

Choose a time

Opens Google Calendar, where you register your details.
Free for qualified teams. No purchase required.