OCT 13, 2026
Moondream builds open-weight vision-language AI models and tools for real-world visual tasks
Moondream builds open-weight vision-language AI models and tools for real-world visual tasks. It offers models for visual reasoning, image captioning, object detection, pointing, and segmentation; a fine-tuning platform (Lens); and Photon, an inference engine for fast deployment across cloud, workstation, server, and edge hardware. Models can be self-hosted or used through hosted APIs, with commercial use supported.