DeepSeek Harness all-in-one
xuxun-oss/dsh-vision-imagen · model-adapters
DeepSeek Harness Plugin: Multimodular Identification and Image Generation for DeepSeek Model Bridge to Google Gemini.
★ 3GitHub stars
0Forks
2026-08-30Last updated
JavaScriptLanguage
MITLicense
Key features
- A DeepSeek sessionDirectly using the regular DeepSeek model (without having to be a visual model), and when looking, drawings, redrawing, the plugin auto-routing to the corresponding Gemini multimodular/living model - the full zero manual switching model.
- Configure (including API Key) to be saved at /.dsh/vision-imagen.json.
- Q: Why do you have to "self-censor" when you're producing the map?
- API Key has only a primary memory and is allergic to page echoes (only 4 bits).
- More complete than the modlensthe modlens are just "an eye for a graphic reading of a pure text model."
Install command
dsh plugin --profile web add github:xuxun-oss/dsh-vision-imagen