DeepSeek MCP Vision Toolkit
shinelon/eyes-for-deepseek · tools
MCP provides image recognition, OCR, UI conversion, error diagnosis, chart understanding and dual image comparison capabilities for plain text models.
★ 5GitHub stars
2Forks
2026-08-13Last updated
TypeScriptLanguage
—License
Key features
- Image parameters are uniformly supported: local file path (jpg / jpeg / png) / public network URL / data: base64.
- 7 visual toolsgeneral image recognition, OCR, UI to code, error diagnosis, chart understanding, data visual analysis, and double image comparison
- Get it immediately after registration, no threshold
- Visual backendSenseTime SenseNova 6.8 Flash-Lite (native multi-modal, "see, think, and do" integration)
Install command
dsh plugin --profile web add github:shinelon/eyes-for-deepseek