This repository publishes prebuilt OCR cache archives for Msty Claw.
The archives are not sidecar runtimes. They mirror the cache layout used by
tesseract-rs so Msty Claw developers and CI can link OCR support without
building Tesseract and Leptonica locally.
Each release is built by .github/workflows/release.yml and uploads:
msty-claw-ocr-runtime-darwin-arm64.tar.gzmsty-claw-ocr-runtime-darwin-x64.tar.gzmsty-claw-ocr-runtime-linux-x64.tar.gzmsty-claw-ocr-runtime-win32-x64.zip- matching
.sha256files
Each archive contains:
cache/leptonica/tesseract/third_party/tessdata/eng.traineddatatessdata/tur.traineddatamanifest.json
Release v0.1.0 also includes a PDF parse fixture archive for QA:
The archive contains generated smoke-test PDFs plus open-source PDFs from OCRmyPDF, Apache PDFBox, libHaru, and Wikimedia Commons. It includes a README, source notes, and per-file checksums for QA handoff.
After downloading both files, verify the archive with:
shasum -a 256 -c msty-claw-pdf-parse-test-pdfs.zip.sha256These assets are versioned with the Rust crates and native library versions used by Msty Claw:
liteparse:2.0.4tesseract-rs:0.2.0- Tesseract:
5.3.4 - Leptonica:
1.84.1
If any of these versions change in Msty Claw, publish a new release here before updating the app build.