dsh-vision-recognizer
CheckedUndeclaredUIv0.1.1Vision
DeepSeek Harness image recognition plugin: keep DeepSeek as the conversation brain while a configurable vision model (15+ providers, OpenAI-compatible + Anthropic) transcribes attached images. Configurable from Settings → Plugins.
Install
dsh plugin --profile web add dsh-vision-recognizerSwap web for your own profile name. First use initializes the profile for you.
Checks
- declares-bundleDeclares dsh.bundle.patch → cordis.patch.yml
- patch-shippedThe patch file ships in the tarball
- patch-parsesPatch parses and mounts 1 plugin row(s)
- rows-resolvableEvery plugin row resolves to a published package
- entry-shippedEntry module ships: lib/index.js
- prebuiltPrebuilt — installs without running build scripts
- client-half-shippedBrowser half ships: client/client.js
What it mounts
insertdsh-vision-recognizerdsh-vision-recognizer
providerIdinnerProviderproviderapiKeymodelbaseURLmaxTokenstimeoutMsmarkerautoLocalOllamafallbackModels
These are the rows the plugin's cordis.patch.yml inserts into your config tree. After installing, dsh --profile <name> --dump-config prints them back.