OPEN-SOURCE SOFTWARE
Built on good work.
The image-to-text tool uses the open-source software below. Their authors retain their copyright. The accompanying licenses describe the permissions and warranty terms.
Local text recognition · Tesseract.js 7.0.0
Tesseract.js 7.0.0 and Tesseract.js-core 7.0.0 run optical character recognition in your browser. Both projects are distributed under the Apache License, Version 2.0. Bell River serves unmodified runtime files from their versioned npm packages.
The tool uses the LSTM recognition engine. The official loader selects a Relaxed SIMD, SIMD, or standard WebAssembly build according to your browser's capabilities. The WebAssembly code is embedded in the matching JavaScript file; only the selected build is loaded.
English & Russian · tessdata_fast
The English and Russian language models come from the official tesseract-ocr/tessdata_fast project, under the Apache License, Version 2.0. They are pinned to commit 87416418657359cb625c412a48b6e1d6d41c29bd.
These compact, integer LSTM models favor responsive recognition on everyday devices. Bell River losslessly compresses the original files with gzip for delivery; the model contents are unchanged. The English download is 1,967,599 bytes and the Russian download is 1,599,713 bytes.
Dependencies & provenance
Tesseract uses Leptonica and other open-source libraries. This software is based in part on the work of the Independent JPEG Group. Additional JavaScript, image-codec, compression, math, and compiler-runtime notices are included below.
- Full third-party copyright and license notices
- Exact file sizes, SHA-256 checksums, source URLs, and package-integrity values
The npm packages were checked against their published SHA-512 integrity values. All runtime and language files are hosted on Bell River. Images and recognized text are processed on your device.
The original authors retain their copyrights. The software and language models are provided without warranty as described in their accompanying licenses.