---
title: 'What is temporal grounding?'
description: 'Temporal grounding finds the start and end time in a video that match a natural-language description.'
url: 'https://videocontextapi.com/glossary/temporal-grounding/'
---

# Temporal grounding

> Temporal grounding is the task of finding the exact start and end time in a video that match a natural-language description, such as "the moment the speaker shows the pricing slide".

## Why temporal grounding is hard

A model can often say *what* happens in a video. It is much harder to say *when* it happens, to the second. Temporal grounding needs the model to connect language to a precise span of frames and audio.

## Where it is used

- **Search.** Return moments, not files.
- **Editing.** Set in and out points for a clip.
- **Agents.** Cite the exact time in a video as evidence.

Video Context returns a start and an end time for every result, so agents and editors can act on it.

Related terms: https://videocontextapi.com/glossary/semantic-video-search/, https://videocontextapi.com/glossary/video-understanding/