Find similar shots

Sometimes you already have a useful shot and want alternatives with related content. ShotAI's similar-shot capability starts from a shot that is already in your library. Use it to discover candidates, then decide which visual qualities are important for the edit.

This is a library-based workflow. It does not establish support for uploading an external screenshot or finished film and identifying the exact original source. Visually similar material is also not proof that two clips are the same recording.

Choose a reference in your library

First find and review a shot through Search your library. Check its source video and range, then decide what should remain similar. You may want the same subject, a related action, a comparable setting, or another option for a particular visual role.

Write down the requirement in visible terms. For example, “another close-up of hands preparing food” gives you a clearer review criterion than “something with the same feeling.” Give the assistant the reference and use that description as the criterion for reviewing candidates.

Request related candidates

The connected-agent workflow exposes a similar-shot operation using an existing library shot as its reference. Once the agent has identified that shot, ask it to find alternatives and preserve the intended collection scope:

Using the library shot we just reviewed as the reference, find similar shots only in “Kitchen footage.” Prefer results from other videos. Return a short list for review, with each source video and its start and end times.

The operation can restrict the collection and exclude the reference video. If your assistant cannot access this operation, use a descriptive semantic search with the same subject and action instead. Do not guess a reference identifier or treat an unrelated image as a library shot.

See Work with the AI assistant for how to make a request explicit about scope, review, and the intended output.

Compare the candidates

Open each promising result and inspect the quality you care about. Similarity can retrieve related content that differs in framing, movement, lighting, or the exact action. It does not automatically select a continuity match for an edit.

Check whether a candidate comes from the right project and whether the source remains accessible. Review the whole returned range: the moment you want may occupy only part of the detected shot.

Decide what to use

Select the candidates that meet the brief and continue with Review and export. Finish precise range choices and continuity decisions in your editing application when needed.

If every candidate misses the required detail, return to a direct description and change one visible constraint at a time. The expected outcome is a reviewable set of alternatives, not a guarantee of exact identity, duplicate detection, source ownership, or a ready-made edit.