Overshoot helps developers build applications that understand live video in real time. It handles video ingestion, model hosting, routing, and inference so developers can connect a camera or screen stream and query frames or video segments through an OpenAI-compatible API. It is designed for robotics, gaming, physical security, sports, automation, and multimodal agent workflows.
Documentation: https://docs.overshoot.ai/ GitHub: https://github.com/Overshoot-ai



