Ds Vision Plugin
Sorwcyra/ds-vision-plugin · model-adapters
Paste images into DeepSeek Harness with a four-model vision race, OCR, and an automatic text bridge.
★ 5GitHub stars
0Forks
2026-08-25Last updated
JavaScriptLanguage
MITLicense
Key features
- Native paste/drop UXusers attach images directly in the Harness Web composer without first saving a path or naming a tool.
- Automatic conversion for Web image attachments
- The proven ds-vision-skill four-model pattern: Agnes 2.5, Agnes 2.0, GLM-4V-Flash, and GLM-4.1V-Thinking-Flash start together
- the first valid response is handed to DeepSeek.
- No GLM 4.6 dependencythe default configuration, routing, and tests contain no glm-4.6v-flash route.
Install command
dsh plugin --profile web add github:Sorwcyra/ds-vision-plugin