OpenCode vision support: automatically saves pasted images and guides non-vision models to recognize them via tool calling. Supports single & multiple images.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).