Process Nepali Hindi and English Documents

Handle mixed Nepali, Hindi, and English documents with per-page language hints.

Send a language_hint per page for mixed-language PDFs
Compare output quality across scripts with confidence scores
Save per-page Markdown for downstream review

Time to complete: ~10 minutes

Prerequisites

A SAGEA_API_KEY exported as an environment variable
A mixed-language file in PDF, PNG, JPG, TIFF, or WEBP format, max 100MB, ideally 300 DPI
Python 3.9+ with the requests package installed
Set up document processing
ARVA OCR model card

Step 1: OCR with a language hint

Pass language_hint alongside the document to bias script recognition per page.

curl -X POST https://api.sagea.space/v1/ocr/process \
  -H "Authorization: Bearer $SAGEA_API_KEY" \
  -F model="arva-ocr" \
  -F [email protected] \
  -F output_format="json" \
  -F language_hint="ne"

ARVA OCR supports 150+ languages with 99.2% accuracy on clean 300 DPI scans.

Step 2: Compare hints per language

Run Hindi and English hints and keep the best-confidence result per page.

curl -X POST https://api.sagea.space/v1/ocr/process \
  -H "Authorization: Bearer $SAGEA_API_KEY" \
  -F model="arva-ocr" \
  -F [email protected] \
  -F output_format="json" \
  -F language_hint="hi"

Each response uses the standard shape with text, pages, confidence, blocks, and model_used set to arva-ocr.

Step 3: Save per-page output for review

Split the best result by page and write Markdown files for reviewers.

from collections import defaultdict
 
by_page = defaultdict(list)
for b in best["blocks"]:
    by_page[b["page"]].append(b)
 
for page, blocks in sorted(by_page.items()):
    blocks.sort(key=lambda x: x.get("confidence", 0), reverse=True)
    with open(f"page-{page}.md", "w") as f:
        f.write(f"# Page {page} (hint {best_hint})\n\n")
        for b in blocks:
            f.write(f"{b['text']}\n\n")
 
print(f"wrote {len(by_page)} page files")

Standard pages cost $1.00 per 1000 pages and complex pages $2.50 per 1000 pages.

Verify

Error	Cause	Fix
`401 Unauthorized`	Missing or invalid API key	Export a valid key as `SAGEA_API_KEY` and retry
`400 unsupported_format`	File type not in PDF, PNG, JPG, TIFF, WEBP	Convert the file to PDF or PNG and retry
`413 Payload Too Large`	File is over 100MB	Compress or split the file, then retry
`429 rate_limit_exceeded`	Too many requests at once	Back off with retries and reduce concurrency