About the Project

OCR Express is a practical OCR service focused on speed, predictable behavior, and a clean developer experience. It accepts uploaded images or remote image URLs, runs OCR using tesseract-wasm in a Node.js runtime, and returns structured per-item results with processing time.

WASM OCR Engine

The OCR route uses a single shared Tesseract WASM engine instance to avoid repeated initialization and keep request latency stable.

Predictable Processing

Requests are queued and processed sequentially, so concurrent traffic does not corrupt OCR state. Inputs are handled in-memory and not persisted as stored uploads.

The Stack

  • Next.js App Router (Node.js runtime for OCR route)
  • tesseract-wasm (WASM OCR engine)
  • TypeScript
  • Tailwind CSS
  • image-js (decode and RGBA preprocessing)

Operational Notes

  • Each OCR result includes either `text` or an item-level `error`.
  • URL image fetches are time-limited to prevent long blocking requests.
  • A queue-based execution model protects engine stability under concurrent traffic.