maxwell-feng/dsh-windows-ocr ↗★ 6
@maxwell-feng/dsh-windows-ocr
dsh plugin: recognize attached images with the built-in Windows OCR engine (Windows.Media.Ocr) and send only the recognized text to the model. Text models never receive image bytes; vision passthrough is opt-in.
AI Analysis
核心用途是免安装第三方 OCR 库，直接利用 Windows 本地能力进行图片文字识别。适合 Windows 平台用户，能有效保护隐私并让纯文本模型具备图片文字处理能力。
Install
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:maxwell-feng/dsh-windows-ocrREADME
Read the full README ↗Configuration
All settings live in the patch row windows-ocr (cordis.patch.yml here) and can be overridden from your profile's cordis.patch.yml:
| Key | Default | Meaning |
|---|---|---|
language | "" | BCP-47 tag for Windows OCR, e.g. zh-Hans, en-US. Empty = user profile languages. |
passthrough | false | false (default): OCR every image. true: genuine vision models receive images untouched. |
ocrScript | bundled lib/ocr.ps1 | Absolute path override for the PowerShell OCR script. |
timeoutMs | 60000 | Per-image OCR timeout. |
maxCacheEntries | 200 | Bound on the per-run OCR cache (keyed by attachment id). |
Example override in ~/.dsh/profiles/web/cordis.patch.yml:
- update:
- id: windows-ocr
config:
language: zh-Hans