Skip to content
All Projects
AI EngineeringProduction

Intelligent Document Processing Platform

AI-Powered Multilingual OCR & Document Intelligence

Converts scanned PDFs and photographed forms into clean, structured Markdown across 10+ languages, including hard scripts like Urdu, Arabic and Amharic, with no manual keying.

FastAPICeleryRedisDoclingRapidOCRTesseract 5TATRGemini 2.0OpenCVNext.js
Problem Statement

Enterprises process thousands of documents monthly. Manual keying averages 6–8 minutes per page and scales linearly with volume, while standard OCR fails on non-Latin scripts and discards tables, headings and reading order.

  • Manual data entry is slow, costly and scales linearly with volume.
  • Standard OCR tools fail on Urdu, Arabic, Amharic and Khmer scripts.
  • Raw OCR loses tables, reading order and headings, flattening the document structure.
Headline Outcomes
Automated

Manual keying eliminated

6–8 min / page

10+

Languages supported

~300 MB RAM

Web process footprint

The Solution

A three-layer pipeline (layout detection, then OCR, then Markdown assembly) behind an async REST API, with Tier-4 routing to Gemini 2.0 Flash when local models hit their accuracy ceiling.

Async five-stage pipeline behind a FastAPI 202-accept pattern keeps the web process lean (~300 MB RAM).

Two Celery queues isolate CPU-bound OCR (prefork) from I/O-bound Gemini calls (gevent).

Docling Heron collapses the DocLayNet taxonomy into 6 canonical labels for accurate structure segmentation.

Tier-4 routing: pages with ≥25% complex-script characters bypass local OCR for best-in-class accuracy.

System Architecture

How the data flows

01

Upload & Probe

REST API, 100 MB limit

02

Language Detect

langdetect + Tesseract OSD

03

Layout Detect

Docling Heron (HuggingFace)

04

OCR + Tables

RapidOCR + Tesseract + TATR

05

Markdown Output

GFM tables + XY-cut order

Result 01

Eliminated expensive manual data-entry workflows at enterprise scale.

Result 02

Unlocked searchable, machine-readable content from legacy multilingual archives.

Result 03

Serves government, legal, research and healthcare digitisation use cases.

Build Something Like ThisServices & PricingMore Case Studies
Further reading

From the blog

AI Engineering

AI Cost Optimization: How We Cut a Document AI Bill by 99%

A practical AI cost optimization guide built on a real case study: how task-based model routing cut one document platform's AI spend by 99% and trims most LLM bills 30 to 50%.

AI CostsLLMCost OptimizationAI Strategy
Read Article12 min read
Taking on new projects · Outside IR35

Have a data pipeline or warehouse problem worth solving?

From messy source data to analytics-ready warehouses that cut cost. Let's scope it. I reply within one business day.

Start a ProjectConnect on LinkedIn