Meow Vision
meimiaoji-creator/meow-vision · model-adapters
Myow-vision is a visual plugin for DeepSeek Harness, which solves the lack of visualization of pure text models.
★ 3GitHub stars
0Forks
2026-08-25Last updated
JavaScriptLanguage
MITLicense
Key features
- The current model cannot process pictures (read image tool will be rejected by dsh's multi-modular door) but you need to feed the model with a map (interception, UI design, error report).
- Models (even text models) call automatically.
- Write a minimum Vue demo in .preview/ src/ desmos/ (render only one component, graphic settings for props/slots)
- The model will receive a text description of the figure and continue its reasoning.
Requirements
- Old version of DSH rc before Cordis API (needs ⁇ 0.0)
- You need to rewrite the dev server entrance for meow preview against demo (e. g. React repository) non-Vue library
Install command
dsh plugin --profile web add github:meimiaoji-creator/meow-vision