# What your agent can do (/docs/agents/what-agents-can-do)

## Make a video

- **"Make a 30-second portrait video about our new app."** — The agent previews first, then submits the same command without `--dry-run`:

```bash
reela create "A 30-second video about our new app" --visual ai-video --layout portrait --duration 30 --dry-run
reela create "A 30-second video about our new app" --visual ai-video --layout portrait --duration 30
```

- **"Use this brief."** — `reela create --prompt-file brief.txt --visual ai-video --dry-run`
- **"Animate this photo."** — `reela create "..." --visual image-video --file original-image:photo.png --dry-run`
- **"Edit my footage."** — `--visual original-clip --file video-source:clip.mp4`, or `--visual original-sync --file video-audio:clip.mp4` to keep its sound.
- **"Make a music video from my song."** — add `--file background-music:track.mp3`. Reela cuts the video to the song with no voice-over; the song needs its lyrics. For background music under narration, the agent describes the music in the prompt instead.
- **"Make a music video."** — add `--music-video`.
- **"Keep these videos together."** — reuse the same `--session "<title>"`.

## Choose quality

- **"What quality levels can I use?"** — `reela tiers list`
- **"Make it at the highest quality."** — add `--tier "<tier-label>"`, copying the full label from `reela tiers list`.

## Use a presenter

- **"Put an avatar in it."** — `reela avatar list`, then `reela create "..." --visual avatar --avatar <avatar-id> --dry-run`

## Check and download

- **"Is it done yet?"** — `reela tasks get <id>`
- **"What's still in progress?"** — `reela tasks list --status processing`
- **"Send me the file."** — `reela tasks download <id>`, or `reela tasks download <id> --out-dir <folder>`

## Apply a brand kit

- **"Add our brand to it."** — `reela brand list`, then `reela brand apply <brand-id> --task <task-id>`
- **"Get the branded version."** — `reela tasks versions <task-id> --origin brand`, then `reela tasks download <task-id> --video <video-id>`

## Account

- **"Is everything set up?"** — `reela status` checks your account and delivery status.
- **"What will this cost?"** — run the `reela create` command with `--dry-run`; the result includes `credits_cost` and `quota_after`.
- **"Which account am I using?"** — `reela whoami`

Agents add `--output json` when they need to read results. Submitting without `--dry-run` can use credits. See [reela create](/docs/cli/reference#create) and [reela tasks](/docs/cli/reference#tasks).