1---
2name: transcribe-captions
3description: Transcribing audio to generate captions in Remotion
4metadata:
5  tags: captions, transcribe, whisper, audio, speech-to-text
6---
7 
8# Transcribing audio
9 
10To transcribe audio to generate captions in Remotion, you can use the [`transcribe()`](https://www.remotion.dev/docs/install-whisper-cpp/transcribe) function from the [`@remotion/install-whisper-cpp`](https://www.remotion.dev/docs/install-whisper-cpp) package.
11 
12## Prerequisites
13 
14First, the @remotion/install-whisper-cpp package needs to be installed.
15If it is not installed, use the following command:
16 
17```bash
18npx remotion add @remotion/install-whisper-cpp
19```
20 
21## Transcribing
22 
23Make a Node.js script to download Whisper.cpp and a model, and transcribe the audio.
24 
25```ts
26import path from "path";
27import {
28  downloadWhisperModel,
29  installWhisperCpp,
30  transcribe,
31  toCaptions,
32} from "@remotion/install-whisper-cpp";
33import fs from "fs";
34 
35const to = path.join(process.cwd(), "whisper.cpp");
36 
37await installWhisperCpp({
38  to,
39  version: "1.5.5",
40});
41 
42await downloadWhisperModel({
43  model: "medium.en",
44  folder: to,
45});
46 
47// Convert the audio to a 16KHz wav file first if needed:
48// import {execSync} from 'child_process';
49// execSync('ffmpeg -i /path/to/audio.mp4 -ar 16000 /path/to/audio.wav -y');
50 
51const whisperCppOutput = await transcribe({
52  model: "medium.en",
53  whisperPath: to,
54  whisperCppVersion: "1.5.5",
55  inputPath: "/path/to/audio123.wav",
56  tokenLevelTimestamps: true,
57});
58 
59// Optional: Apply our recommended postprocessing
60const { captions } = toCaptions({
61  whisperCppOutput,
62});
63 
64// Write it to the public/ folder so it can be fetched from Remotion
65fs.writeFileSync("captions123.json", JSON.stringify(captions, null, 2));
66```
67 
68Transcribe each clip individually and create multiple JSON files.
69 
70See [Displaying captions](display-captions.md) for how to display the captions in Remotion.
71