> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pureframe.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Overview

> Find any moment in your video library using text, an image, or both.

`POST /v1/search` finds timestamped moments in your video library. Send a text description, a reference image, or a public image URL — results are ranked by relevance and grouped by video.

## Input types

| Input      | How to send            | Example                         |
| ---------- | ---------------------- | ------------------------------- |
| Text query | `query` form field     | `"person waving at the camera"` |
| Image file | `image` file upload    | Upload a JPEG or PNG            |
| Image URL  | `image_url` form field | `"https://example.com/ref.jpg"` |

At least one input is required. You cannot send both `image` and `image_url` in the same request. A text-only query also requires `collection_id` — image search can run across your whole library, but text search needs a collection to scope against.

## What gets searched

Every video is indexed across up to four signal types, and a query can match any of them:

| Type         | What it matches                                                          | Search page                                                                 |
| ------------ | ------------------------------------------------------------------------ | --------------------------------------------------------------------------- |
| `frame`      | Visual content — what's on screen                                        | [Image search](/search/image), [Combined queries](/search/combined-queries) |
| `transcript` | Spoken words, transcribed                                                | [Transcript search](/search/transcript)                                     |
| `scene`      | Auto-detected scene labels (e.g. outdoor, whiteboard) at shot boundaries | [Combined queries](/search/combined-queries)                                |
| `ocr`        | Text visible in frame — signage, slides, captions                        | [Combined queries](/search/combined-queries)                                |

Each result segment's `type` field tells you which one matched. See [Search Results](/search/search-results) for the full response shape.

## Scoping and pagination

Pass `collection_id` to limit results to one collection — faster and more precise than searching your whole library. Results are paginated with `page` and `per_page` (max 50).

```bash theme={null}
curl -X POST https://api.pureframe.ai/v1/search \
  -H "Authorization: Bearer pf_..." \
  -F "query=product demo" \
  -F "collection_id=col_abc123" \
  -F "per_page=20"
```

## Credits and limits

Each search request consumes 1 credit on Pay as you go accounts (see [Usage and Billing](/production/usage-and-billing)). Free accounts get a daily quota of 30 searches. Requests are rate-limited to 30/minute per API key — see [Rate Limits](/production/rate-limits).

<CardGroup cols={2}>
  <Card title="Text Search" icon="font" href="/search/text">
    Natural language queries against visual and spoken content.
  </Card>

  <Card title="Image Search" icon="image" href="/search/image">
    Find visually similar frames from a reference image.
  </Card>

  <Card title="Transcript Search" icon="closed-captioning" href="/search/transcript">
    Search only what was said, not what's shown.
  </Card>

  <Card title="Combined Queries" icon="layer-group" href="/search/combined-queries">
    How modes merge, and the `modes`/`labels` parameters.
  </Card>
</CardGroup>