llm-vision

AI Video Editor Pipeline with Vision LLM Models

Claude Code skill — 给无视觉能力的 LLM（DeepSeek/o1/o3）外挂看图能力。Vision proxy for text-only LLMs, supports OpenAI/Anthropic/DashScope/Qwen3-VL.

A lightweight Model Context Protocol (MCP) server for retrieving local and remote images for LLM vision models, featuring metadata extraction and configurable availability retries.

DSH plugin: vision_read — route image reading to a dedicated vision model (e.g. Kimi K3) so text-only agents can see images

Alt'Ollama: The app for generating alt-text for image/s using LLMs with image processing support.

LLM Vision integration for Home Assistant using Google Gemini

Home Assistant Card to display the LLM Vision Timeline