dsh-vision-toolkit
dsh-vision-toolkit adds a real vision toolbox to text-only AI agents: paste an image for recognition, multi-image Q&A, screenshot-to-UI restoration, and pixel-level diffing.
View anionex/dsh-vision-toolkit on GitHub- GitHub stars
- 946
- Forks
- —
- Language
- TypeScript
Project details
- Repository
- anionex/dsh-vision-toolkit
- Licence
- MIT
- Primary language
- TypeScript
- Catalog status
- Available
- Catalog check
- 2026-11-13
Repository description
[dsh]为纯文本模型设计更强大的视觉工具箱：一行安装使用、粘贴图片直接识别、多张图片问答、截图到前端UI 还原等｜DeepSeek Harness-native integration for agent-vision-toolkit: image Q&A, long-screenshot OCR, UI restoration, grounding, pixel diff, Artifacts, and Web UI.
Explore by topic
More tools in Knowledge & documents
- agent-vision-toolkit
anionex · MIT
Gives text-only AI coding agents the ability to understand images by handling image Q&A, screenshot OCR, and UI recognition outside the model itself.
- ai-engineering-from-scratch
rohitg00 · MIT
A free, open-source course that walks you through building modern AI systems from scratch, 511 lessons spanning LLMs, RAG, agents and transformers.
- headroom
headroomlabs-ai · Apache-2.0
Headroom shrinks everything an AI agent reads (logs, files, RAG chunks, tool output) before sending it to the LLM, cutting token costs without changing the answers.
- nuphus-mcp
mrpulor-gh · MIT
Gives AI agents actual computer-use capability: they can see your screen and control windows, mouse, keyboard, and Chrome through a lightweight local server.