# OCR engine licenses

Everything under `public/ocr/` is served from the app origin. The editor does not fetch OCR weights, WASM, or the worker from a CDN.

## tesseract.js — Apache-2.0

`worker.min.js` is the browser worker from [tesseract.js](https://github.com/naptha/tesseract.js) (Apache License 2.0).

## tesseract.js-core — Apache-2.0

The WebAssembly builds and their JS glue (`tesseract-core*.wasm`, `tesseract-core*.wasm.js`, including the `simd`, `relaxedsimd`, and `lstm` variants) come from [tesseract.js-core](https://github.com/naptha/tesseract.js-core) (Apache License 2.0). That package compiles the Tesseract OCR engine (also Apache-2.0) to WASM.

## Tesseract traineddata — Apache-2.0

Language models for OCR:

| File | Language |
| :--- | :--- |
| `eng.traineddata.gz` | English |
| `spa.traineddata.gz` | Spanish |
| `fra.traineddata.gz` | French |
| `deu.traineddata.gz` | German |

These are Tesseract `tessdata_fast` models (`eng`, `spa`, `fra`, `deu`), Apache License 2.0, from the [tesseract-ocr/tessdata_fast](https://github.com/tesseract-ocr/tessdata_fast) project (and the matching `@tesseract.js-data` packs, same license).
