uknowmyface/locallens

Extracts text from screenshots, clipboard images and image files using macOS's built-in Vision framework, so a text-only model can read an image that never leaves the Mac.

Comparable plugins route the image to a vision model; LocalLens calls Apple's Vision framework, the engine behind Live Text, directly, which removes keys, quotas and uploads at the cost of reading text rather than understanding pictures. It handles screen captures, clipboard images and image files, and the README is explicit about the trade: it is OCR, so it will not describe a scene, and it does not work off macOS. The security notes and SECURITY.md are part of the repository because the value proposition is entirely about the image staying local, and it is positioned against visionDS, dsh-vision-router and dsh-vision-sidecar, which all send the image to a model.

Vision & Multimodal ★ 1
View on GitHub ↗

Install

dsh plugin --profile web add github:uknowmyface/locallens

The README's install line, from GitHub; no npm package is documented. The plugin ships its own cordis.patch.yml declared through dsh.bundle.patch, so dsh plugin add installs and registers it in one step.

Compatibility

macOS only, version 13 (Ventura) or later; below 13 the recognitionLanguages option is unavailable, the languages setting is ignored and Chinese accuracy degrades. It needs no API key, quota or network access. The README states plainly that it is an unofficial community project, not affiliated with DeepSeek or Apple.

Details

Alternatives

fieldnote-ops/frameevidence · maxwell-feng/dsh-windows-ocr · maxwell-feng/dsh-tesseract-ocr

More plugins in Vision & Multimodal

Browse more in Vision & Multimodal

Guides for Vision & Multimodal plugins