FuzzySoul/dsh-free-vision ↗★ 1
dsh-free-vision
Free vision plugin for DeepSeek Harness (dsh): image understanding for text-only models with free-tier providers (Qwen3-VL-Flash / DeepSeek-OCR / Doubao). 免费视觉插件：纯文本模型看图能力，优先免费模型（通义千问 / 硅基流动 / 豆包）。
AI 분석
核心用途是低成本为纯文本模型扩展看图能力。适合预算有限、但需要模型处理和分析图片内容的用户。
설치
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:FuzzySoul/dsh-free-visionConfiguration / 配置
- id: free-vision
name: 'dsh-free-vision'
config:
apiKey: 'sk-xxxx' # optional: falls back to the provider env var
modelProvider: qwen # qwen | volcengine | siliconflow | zhipu | hunyuan | custom
modelName: qwen3-vl-flash # optional model override
toolName: image_understand # tool public name (rename if it collides)
maxTokens: 8192
temperature: 0.7
multiCrop: true
toolCallTimeoutMs: 200000
lumaEnv: {} # extra env vars for the vision engine
Or just set the matching environment variable (e.g. DASHSCOPE_API_KEY).
也可以只设置对应的环境变量（如 DASHSCOPE_API_KEY）。
Free API keys / 免费 Key 申请
| Provider | Where to get a free key |
|---|---|
| qwen | 阿里云百炼 bailian.console.aliyun.com — 开通即送免费额度，模型选择 qwen3-vl-flash（限免） |
| volcengine | 火山引擎 volcengine.com — 豆包模型新用户送免费 token（20万起，可申请 50万） |
| siliconflow | 硅基流动 siliconflow.cn — DeepSeek-OCR 免费调用 |
Usage / 用法
The model calls image_understand with:
image_source(required): local file path, HTTP(S) URL, or data URI (PNG/JPG/WebP/GIF, ≤10MB)prompt(required): the question about the image — works in English or Chinesetask_type(optional):auto | general | ocr | ui | debug | describe