vss-summarize-video

Automated video summarization routes clips through a human-in-the-loop LVS workflow, enabling scenario-driven event extraction for incident reporting. You provide a scenario and target events, and the workflow outputs structured summaries.

Video summaries are always produced even when LVS is unavailable by automatically falling back to a VLM. This fallback path ensures continuity for your video analytics workflow without manual intervention.

Video event extraction supports configurable chunking, frames, and objects of interest through scenario-driven inputs. You define the specific scenarios and events to surface, and the summarization logic processes the video accordingly.

LVS provides human-in-the-loop guided video summarization for high-quality, explainable outputs, while the VLM fallback automatically processes clips when LVS is unavailable. The routing logic prioritizes LVS for readiness and falls back to VLM for continuity.

Video summarization supports structured output handling tailored for safety, security, and incident reporting workflows. It satisfies integration requirements for model selection and LVS readiness while extracting scenario-driven events from the provided clips.