Drop2Hear · Beta
Open source notices
This app builds on open-source tools. The list below identifies major components, OCR data, and the English speech model.
License information
Licenses belong to their respective authors. See each project for its terms. The build also publishes a generated third-party license file for bundled npm dependencies.
Model and data files
English, Spanish, French, German, Hindi, and Portuguese OCR trained data are bundled with the app under the upstream Apache-2.0 license, published alongside the data. Natural Audio downloads the Kokoro-82M-v1.0-ONNX model from Hugging Face when needed. The model card identifies an Apache-2.0 license.
Attribution limits
This page is an attribution aid, not a replacement for the full license texts or a comprehensive legal review. Additional transitive packages appear in the generated license file.
Components and licenses
| Component | Purpose | License | Source |
|---|---|---|---|
| Angular | User interface, routing and PWA support | MIT | Project source |
| PDF.js (pdfjs-dist) | PDF text extraction and rendering | Apache-2.0 | Project source |
| Tesseract.js | Browser-based OCR | Apache-2.0 | Project source |
| Multilingual OCR trained data (naptha/tessdata) | English, Spanish, French, German, Hindi, and Portuguese text recognition data | Apache-2.0 | Project source |
| Mammoth | DOCX text extraction | BSD-2-Clause | Project source |
| kokoro-js | Natural Audio integration | Apache-2.0 | Project source |
| Transformers.js | Browser model loading and inference | Apache-2.0 | Project source |
| ONNX Runtime Web | Browser model inference runtime | MIT | Project source |
| Kokoro-82M-v1.0-ONNX model | English speech model | Apache-2.0 | Project source |