Harvey-Will/dsh-vision-analysis

README 摘要

[] English · 中文 ✨ Why DSH Vision Analysis? Your text-only agent can finally "see" — without swapping your model, without leaking image bytes into the conversation, and without depending on a single vendor. - 8 analysis modes out of the box — describe , ocr , ui-review , chart-data , object-detect , compare , code-gen , debug — each with a tuned instruction template. - Any vision endpoint — OpenAI chat/completions or Anthropic messages wire formats. MiMo, Step, SiliconFlow, OpenRouter, Gemini (OpenAI-compat), GPT-4o, Claude, Qwen-VL, or a local Ollama / LM Studio / vLLM. - Any input — absolute local path, http(s) URL, or base64 data: URL; up to 4 images per call with built-in comparison. - Privacy-first by design — image bytes never enter the session log or reach your main model; only the vision model's text comes back. The debug report never reveals your API key (fully masked). - Live configuration — edit endpoint, model, and per-mode tuning from Settings → 插件配置 with secrets masked. - Web UI guidance — when your active model can't take images, a composer hint tells you the reliable path: save locally → send the path → analyze image parses it (native image routes stay untouched). - …

分类