Marketplace / recon-scanner
Recon Scannerrecon-scannerv1.4.0

PWA document scanner with Tesseract.js OCR plus inline translation. Offline-first.

WebsiteTSPWAedition of 100
Live siteDocsChangelog
live demo
Recon Scanner
forkit.run/recon-scanner
Live website · runs in a sandbox, no installLaunch live demo →
README.md

A complete, installable document scanner that runs entirely in the browser: scan, OCR and translate without a backend.

Overview

Recon Scanner is a production PWA. Users point their camera at a document (or drop a PDF/image), it extracts text with Tesseract.js and translates inline. Works offline after first load and installs to the home screen.

Quick start

Clone your fork, then bun install and bun run dev. A static build ships with bun run build. Deploy the dist/ folder to any static host (Vercel, Netlify, Cloudflare Pages).

Configuration

Set the translation provider in .env (TRANSLATE_API). Everything else is zero-config. No server, no database.

Requirements

  • Bun 1.x (build only)
  • Any static host
  • Optional: translation API key

Stack

TSPWA

Pinned dependencies, no lock-in to the seller's accounts. Deploy scripts included.

How it works, end to end

1. ingest → stream events from source
2. process → parse, score, decide
3. act → submit transaction / render output
4. log → record outcome + metrics

What it has

  • Tesseract.js OCR
  • Inline translation
  • Offline-first PWA
  • Camera capture
  • Installable to home screen

Can be improved

  • No test suite
  • Translation provider hardcoded