PROJECT TOPICS
INSTALL REFERENCE
dsh plugin --profile web add github:zoahdev/dsh-vision
该命令指向仓库当前默认分支；尚无绑定当前 commit 的完整验证结果。
PROJECT README
Give DeepSeek Harness eyes: a
vision_analyzetool that sends an image (local path or URL) to an OpenAI-compatible vision model and returns a text answer.
English | 中文
dsh plugin --profile web add dsh-visionSet apiKey (or OPENAI_API_KEY), or point baseUrl at a local OpenAI-compatible
endpoint such as Ollama (http://localhost:11434) with a vision model.
| Parameter | Type | Description |
|---|---|---|
image |
string | Image path or https/data: URL |
question |
string | What to ask (default: "Describe this image in detail.") |
| Key | Type | Default | Description |
|---|---|---|---|
apiKey |
string | (env) | OpenAI-compatible API key |
baseUrl |
string | https://api.openai.com |
API base URL |
model |
string | gpt-4o-mini |
Vision model id |
timeoutMs |
number | 120000 |
Request timeout |
给 DeepSeek Harness 装上眼睛：
vision_analyze工具把图片（本地路径或 URL） 发给 OpenAI 兼容的视觉模型，返回文字描述。
dsh plugin --profile web add dsh-vision配置 apiKey（或 OPENAI_API_KEY），或把 baseUrl 指向本地 OpenAI 兼容端点
（如 Ollama http://localhost:11434 + llava）。
CLASSIFICATION EVIDENCE
系统优先读取 GitHub Topics，再与站内分类词典和词根规则比对。当前命中： image、multimodal、vision。
liustack
The first vision plugin for DeepSeek Harness, and the vision bridge for every text-only coding agent. Paste an image, get structured JSON evidence (OCR, layout, semantics). | 全网最强 DeepSeek Harness 外挂视觉插件，为 DeepSeek、GLM 等纯文本模型外挂视觉能力，粘贴图片即得结构化 JSON 证据（OCR、版面、语义）。
Anionex
为纯文本模型"看图“设计更好的视觉工具箱和技能，支持多图理解，图片问答，前端UI还原、GUI 自动化等，并可选无缝接入多个主流agent，直接识别粘贴图片｜ A vision toolkit and skill designed for text-only llms — image Q&A, long-screenshot OCR, frontend UI restoration, and GUI automation, with optional seamless integration for Codex, Claude Code, Pi, Oh My Pi, and OpenCode
ysr666
Eyes for text-only DeepSeek Harness agents: built-in free vision chain (no key) + pixel-level vision tools (Q&A, grounding, crop, pixel diff, colors, OCR, SVG trace, cutout, screenshots). One-command install, no Python, image turns work like ordinary tool-calling turns.