DSH Auto Vision
NormanFxxkingRockwell/dsh-auto-vision · model-adapters
DeepSeek Harness Visual Bridge: Automatically discovers your configured multi-modular model, installs a viewing tool for the pure main text model and recognizes that the result returns in plain text.
★ 2GitHub stars
0Forks
2026-08-25Last updated
JavaScriptLanguage
MITLicense
Key features
- Fill the pure text master model in DeepSeek Harness with eyes: Autodiscover your configured multi-modular model, a command with a view tool, and photo recognition returns in plain text.
- Autohidden read imagethe session of a pure main text model hides the inevitable failure read image, leaving the model to go only by viewing and not crashing once before changing paths
- Autodiscover modelDefault zero configuration.
- You give it a path to the photo file (or let the model access a picture file) and it automatically adjusts the view.
- Plugin lists all registered models that state the image pattern as candidates, sorting them in "Focus price" by "provider ⁇ the rest of the registered provider that you configure for settings"
Requirements
- View, argumentfile path (single chart) or file paths (multi-graph path array with a maximum of 10 requests at a time, described by number), instruation (optional, identification requirements)
- Common set of attachment lines and size limits with read image
Install command
dsh plugin --profile web add github:NormanFxxkingRockwell/dsh-auto-vision