Documents

SmartID AI

IDs, driver's licenses and passports read and structured automatically to pre-fill forms.

The problem

Onboarding flows ask people to type data that is already printed on their ID.

How it works

Google Cloud Vision reads the image, and an LLM maps the noisy OCR text to clean fields — names, document number, issue and expiry dates. Jobs run asynchronously behind a FastAPI service.

Pipeline

01Image
02Validation
03OCR
04LLM field mapping
05Form data

Highlights

Handles ID cards, driver's licenses and passports
Async job API: upload, status, result
Processing metrics per document
Deployed on Railway

Stack

Google Cloud VisionOpenAILangGraphFastAPIRailway

Want something like this for your company?

I build production versions of these systems on your data and your infrastructure.

Related service: Document processing

Book a call