image-recognition

OpenCode vision support: automatically saves pasted images and guides non-vision models to recognize them via tool calling. Supports single & multiple images.

Containerized self-hosted REST API for vision classification, utilizing Hugging Face transformers.