> Documentation index: https://api.veed.io/llms.txt. Fetch it to find every other page, as Markdown.

# Submit a reframe job

Model: Reframe
Endpoint: `POST /v1/reframe`

Start an asynchronous job that turns a landscape video into a vertical 9:16 video. For each shot, the model picks a layout from the frames, the faces, who is speaking and the transcript: it follows the speaker, tracks a moving subject, fits a wide view, or stacks two panels (two people talking, or a speaker and their slides). Captions can be burned in.

**Inputs**

- `video_url` — public URL of a landscape video with both a video and an audio track, up to 10 minutes and 4 GB. The active speaker is found by voice
- `captions` — optional; the style of burned-in captions, or `none` (the default). Captions are transcribed from the speech, with the language detected automatically
- `accent_color` — optional; the colour of the spoken word, as `#rrggbb`. Only `highlight` (default `#ffd43b`) and `pill` (default `#7b61ff`) use it

**Caption styles**

- `clean` — white text with a thin outline and a soft shadow
- `boxed` — black text on a white box behind each line
- `highlight` — short cues; the spoken word is in the accent colour and each cue pops in
- `pill` — all caps, two or three words per cue; the spoken word sits on a pill in the accent colour

Captions are centred and placed where the layout leaves room: on the seam between stacked panels, in the bands around a fitted view, or at about 60% of the height away from faces. They stay clear of the top 15% and the bottom 35% of the frame, where platforms draw their controls.

**Output**

An H.264 MP4 at the source height, capped at 1920 and never upscaled (1080p gives 606×1080, 4K gives 1080×1920). It keeps the source audio and timing, so it is as long as the input. The same video can be framed differently on different runs.

**What happens next**

The job is **accepted immediately** — you get `202 Accepted` with a `job_id` and status `PROCESSING`. Processing takes minutes. Poll `GET /v1/reframe/{job_id}` until the job is `COMPLETED` or `FAILED`.

## Request body

- `accent_color` · string · optional — Colour of the spoken word, as #rrggbb. Only the highlight and pill styles use it. Omit it for the style default.
- `captions` · enum · optional — Style of the burned-in captions, or none for no captions. Accepts `none`, `clean`, `boxed`, `highlight`, `pill`. Defaults to `"none"`.
- `video_url` · url · required — URL of a landscape video with both a video and an audio track, up to 10 minutes and 4 GB.

## Parameters

- `X-Veed-Store-IO` (header) — Set to `0` to not store this request's and response's bodies. They are then not available in your request logs for debugging.
- `X-Veed-Media-Expiration-Seconds` (header) — Number of seconds before the media URLs returned for this request expire. A value above the maximum is capped rather than rejected.

## Example request

```bash
curl -X POST "https://api.veed.io/v1/reframe" \
  -H "Authorization: Bearer $VEED_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "video_url": "https://static-assets.veed.io/api-examples/reframe-input.mp4"
}'
```

## Example response (202)

```json
{
  "data": {
    "job_id": "123e4567-e89b-12d3-a456-426614174000",
    "status": "PROCESSING",
    "credits_estimated": 50
  }
}
```

Rendered page: https://api.veed.io/docs/api/post-v1-reframe
