# Captions

PointerCam writes captions from your voice on your own computer, timed word by word. Edit any line, style them, and light each word as it is spoken.

Where to find it: Video editor › Pro › Audio & Captions › Captions

## Make captions

1. In Simple, find **Captions** under Done for you and press **Transcribe**. The first time, PointerCam downloads the speech model into your browser or app. After that it runs at once.
2. Wait while it works. It keeps going if you leave the tab. One video is transcribed at a time.
3. Read the transcript in **Audio & Captions › Captions**. Click a line to jump there, and type to correct it.

No voice on the take? Press **Import** to bring in an .srt or .vtt file written elsewhere, or **Write a line at the playhead**.

## How they look

In Simple, the rows under **Captions** set the **Size** (**S**, **M** or **L**), the **Plate** (**Dark**, **Light**, or **Glass**: dark words on frosted glass), the **Words** (the whole **Line**, or **One at a time**) and the **Place** (**Bottom**, **Middle** or **Top**). Pro has the numbers:

- **Size** and **Width**: the type size and how wide a line may run, as a share of the picture.
- The same type, background, shadow and arrival as callouts.
- **Placement**: **Bottom**, **Middle** or **Top**, with a margin, or drag the words anywhere on the preview.

## One word at a time

**One word at a time** shows each word alone as it is said, coming into focus, as short-form videos do. It needs the word timings a transcript made here has; the subtitle file keeps whole lines.

## Light the words as they are spoken

- **Off**: A plain caption block.
- **Tint**: Colours the word being said.
- **Select**: Sets the word being said in white on a coloured block, which moves along the line from word to word.
- **Read-along**: Dims the words not spoken yet. **Unspoken opacity** sets how faint they are.

## Keep them out of the picture

Switch off **Burn into the video** and the words stay out of the picture. They still export as a subtitle file, which a viewer can switch off and a search engine can read. See [Save a subtitle file](https://pointercam.com/help/subtitle-file).

> **Tip:** **Remove ums and uhs** tidies the transcript in one press.

## Engines

- **On this machine**: A Whisper speech model installed in this browser or app. You can remove a model to free room.
- **Connected**: A speech model you already run on your own computer, reached at its local server address.

> **Note:** Both engines run on your own computer. Your voice is never uploaded.

## Read next

- [Save a subtitle file for YouTube](https://pointercam.com/help/subtitle-file)
- [Volume, muting and the mix](https://pointercam.com/help/audio-mix)
- [Add text, arrows and boxes](https://pointercam.com/help/callouts)

Source: https://pointercam.com/help/captions