Your documents stay local
The desktop application runs its own authenticated local engine. It does not require hosted document storage, analytics, or a remote-model fallback.
Local-first · open source · source read-only
Doc Evidence inventories, extracts, searches, and compares documents on your computer. Every result stays connected to its source file and page, while machine observations remain distinct from reviewed facts.
The desktop application runs its own authenticated local engine. It does not require hosted document storage, analytics, or a remote-model fallback.
Collections are read-only. Content is identified by SHA-256, paths remain aliases, and generated artifacts live separately from source bytes.
Extracted text and candidate observations retain file and page provenance. Automated agreement never masquerades as human confirmation.
One durable local library
Track immutable content across one or more explicitly authorized folders without confusing renamed paths for new evidence.
Use bundled Poppler, Tesseract, and OCRmyPDF tools offline, with exact extractor versions and cache identity retained.
Inspect source pages beside native text and OCR output, collapse exact matches, and focus on numeric or textual disagreements.
Queue, cancel, retry, and recover extraction work while keeping attempts, warnings, timing, and resulting artifacts inspectable.
Signed desktop builds
The first signed macOS and Windows release has passed exact installed-artifact validation. Both platforms will be published together after final maintainer approval.
No system Python, Homebrew, development tools, or internet connection is required after installation.
Built in public
Doc Evidence is Apache-2.0 licensed. Its architecture, data contracts, desktop packaging manifests, release validation, and synthetic fixtures are available in the public repository.
Explore the source on GitHub →