SnapScribe AI

Document intelligence engine

AI-Powered Document Intelligence

Turn images into
structured data

SnapScribe AI is a modular document intelligence engine. OCR is the reusable core. AI interpretation is optional, never a replacement.

Features

Three extraction modes built on a clean, extensible architecture.

Raw Text Extraction
Core OCR engine powered by Tesseract. Upload any image and get clean raw text with line-by-line segmentation. No AI required.
Structured Documents
Automatically classify and structure receipts, invoices, resumes, and notes into typed JSON fields using Groq LLM.
Smart Analyze Mode
Paid feature. Fully LLM-driven interpretation for unknown or messy documents. Password-protected; contact the developer for access.
Modular Architecture
OCR and AI are cleanly separated. The reusable OCR core can power PDF processing, batch pipelines, and enterprise systems.
Hybrid Pipeline
Deterministic OCR first, optional AI second. Reliable text extraction with intelligent structuring only when you need it.

About SnapScribe AI

SnapScribe AI was built by Jhan Cedric Antonio to demonstrate real AI engineering design: a reusable OCR core layer, optional LLM interpretation, and clean API separation between deterministic and intelligent processing.

Unlike simple OCR apps, SnapScribe is architected as a foundation for scalable document processing, ready to extend into PDF parsing, batch pipelines, video OCR, and enterprise document intelligence systems.

The stack is FastAPI + Tesseract on the backend and Next.js on the frontend. Stateless, per-request processing with no database or authentication required for the MVP.

01

Upload

Drop a JPEG, PNG, or WebP image up to 10MB.

02

Choose Mode

Raw OCR, structured document, or smart analyze.

03

Extract

Processing runs per request. No account needed.

04

Export

View raw text, JSON, or a field table. Copy or download.

Other Works

More full-stack AI projects from Jhan Cedric Antonio.

OLJfinder preview
OLJfinder
A full-stack job discovery tool that scrapes publicly listed developer roles on OnlineJobs.ph, caches them for fast browsing, and layers AI-powered classification plus rich client-side filters so developers find relevant posts faster and apply on OnlineJobs.ph. Unofficial personal project. Listings stay on OnlineJobs.ph, no personal data stored, cache-first API on Render with Next.js on Vercel.
PlanForge preview
PlanForge
A portfolio-grade full-stack demo built to show product-minded engineering, not just LLM integration. Users paste a project idea and deadline; AI generates a structured plan as JSON. Everything after that is fully editable: tasks, dates, order, and timeline placement, with no re-prompting. The app doubles as a polished case-study site with a live interactive demo.
Support & Contact
Interested in sponsoring this work, collaborating, or hiring for your next AI product? Reach out through any channel below.