Thai & Hebrew Real-World Document Collection
Budget: ₹12,500 – ₹37,500 INR
We have a requirement for real-world documents in the Thai and Hebrew languages. Please find the guidelines below:
### Real-World Document Collection Guidelines (Thai & Hebrew)
Please collect **real-world documents** in **Thai** and **Hebrew** following the requirements below:
**Languages**
Thai
Hebrew
Document Requirements**
* Only **dense documents** containing **more than 300 words** are accepted.
**Document Format Distribution**
* **50% Printed** (scanned or mobile-captured printed documents)
* **25% Handwritten** (e.g., handwritten letters, application forms, doctor's notes)
* **25% Digital** (native digital PDFs or machine-generated documents)
**Priority Document Types**
* Loans / Mortgage
* Newspapers
* Legal Documents
* Contract Documents
* Financial & Environmental Reports
* Tax Documents
* Handwritten Doctor's Notes
* Handwritten Applications
* Receipts
* Invoices
* Certificate of Analysis
* Insurance Documents
* Hospital Claims
* Patient Forms
* Checks & Deposit Slips
**Layout Distribution**
The collected documents should include a mix of:
* Text-only pages (26%)
* Tables (20%)
* Key-Value Areas (15%)
* Clickable elements (15%)
* Signatures (12%)
* Figures/Images (12%)
All documents should be in single page.
Use only dense documents that have more than 200 words.
All documents should be real except PII information. While reviewing through auto script, if we find any errors, the entire document needs to be changed.
Every document should be variant with layout elements mentioned in the SOP.
At least 1 or 2 layout elements should be present in each document.
Scanned document variation : Source real documents from open sources and apply variation while scanning, such as coffee stains, wrinkles, dirt marks, etc., to ensure diversity.
Up to 15% Bilingual content is acceptable.
Documents from 2016 to 2026 April are acceptable.
Documents should be in the standard A4 size with 300 DPI.
File size should be less than 10 MB.
Single-page documents should be provided only in JPG or PNG format.
### Real-World Document Collection Guidelines (Thai & Hebrew)
Please collect **real-world documents** in **Thai** and **Hebrew** following the requirements below:
**Languages**
Thai
Hebrew
Document Requirements**
* Only **dense documents** containing **more than 300 words** are accepted.
**Document Format Distribution**
* **50% Printed** (scanned or mobile-captured printed documents)
* **25% Handwritten** (e.g., handwritten letters, application forms, doctor's notes)
* **25% Digital** (native digital PDFs or machine-generated documents)
**Priority Document Types**
* Loans / Mortgage
* Newspapers
* Legal Documents
* Contract Documents
* Financial & Environmental Reports
* Tax Documents
* Handwritten Doctor's Notes
* Handwritten Applications
* Receipts
* Invoices
* Certificate of Analysis
* Insurance Documents
* Hospital Claims
* Patient Forms
* Checks & Deposit Slips
**Layout Distribution**
The collected documents should include a mix of:
* Text-only pages (26%)
* Tables (20%)
* Key-Value Areas (15%)
* Clickable elements (15%)
* Signatures (12%)
* Figures/Images (12%)
All documents should be in single page.
Use only dense documents that have more than 200 words.
All documents should be real except PII information. While reviewing through auto script, if we find any errors, the entire document needs to be changed.
Every document should be variant with layout elements mentioned in the SOP.
At least 1 or 2 layout elements should be present in each document.
Scanned document variation : Source real documents from open sources and apply variation while scanning, such as coffee stains, wrinkles, dirt marks, etc., to ensure diversity.
Up to 15% Bilingual content is acceptable.
Documents from 2016 to 2026 April are acceptable.
Documents should be in the standard A4 size with 300 DPI.
File size should be less than 10 MB.
Single-page documents should be provided only in JPG or PNG format.