limccn/deepseek-vl-support

README excerpt

deepseek-vl-support 中文说明 → docs/README.zh-CN.md What this does Some AI models (such as DeepSeek) are text-only: they can read your files, but they cannot look at pictures. Screenshots of errors, UI mockups, charts — these are invisible to such models. This small tool gives them "eyes". Once installed, whenever the model tries to read a picture, the tool sends that picture to a vision service of your choice (Moonshot, OpenRouter, SiliconFlow, Ollama …), receives a detailed text description, and hands that description to the model. The model then works from the description, as if it could see the picture. No model settings to change, no extra configuration files to write — it works automatically after a one-time setup. One command to install, one command to remove. MIT licensed. Who this is for - You use a text-only model (such as DeepSeek) in any of the supported agents: native — Claude Code, Codex, OpenCode; skill-based — Trae, Pi Coding Agent, DeepSeek Harness; Agent Plugins clients — GitHub Copilot, Cursor, Kiro, OpenClaw, Hermes Agent, VS Code, ChatGPT & Codex, Grok Bot, NanoClaw, and other spec-compliant agents. - You want that model to understand pictures: error screenshots, U…

Category