get-youtube-transcript

Fetch a YouTube transcript / subtitles from the command line or in Node.js — free, no API key. Works in 2026: it handles YouTube's PoToken (BotGuard) requirement that breaks most older transcript tools.

npx get-youtube-transcript https://www.youtube.com/watch?v=jNQXAC9IVRw

All right, so here we are, in front of the elephants...

Why another one?

Most free YouTube transcript libraries stopped working in 2025–2026: YouTube now returns an empty response to caption requests that lack a valid PoToken bound to the video. This tool solves the BotGuard challenge and fetches the real captions, so it keeps working where naive scrapers return nothing.

Install

npm install -g get-youtube-transcript
# or run without installing:
npx get-youtube-transcript <video>

CLI usage

youtube-transcript <video-url-or-id> [--lang en,es] [--format text|json|srt]

youtube-transcript jNQXAC9IVRw
youtube-transcript jNQXAC9IVRw --format json
youtube-transcript jNQXAC9IVRw --lang es --format srt

Use in code

import { getTranscript } from 'get-youtube-transcript';

const { text, segments, language } = await getTranscript('jNQXAC9IVRw', { languages: ['en'] });
console.log(text);
// segments: [{ start, duration, text }, ...]

Scope & limits

This is a single-video, best-effort tool that runs from your own IP. That's perfect for scripts, notebooks, and low-volume use. It is not built for:

• Bulk extraction (hundreds/thousands of videos)
• Reliability at scale — YouTube blocks datacenter IPs and rate-limits aggressively, so unattended/server use will get throttled
• SRT/VTT/RAG-chunk output, 100+ language fallback, or guaranteed uptime

Need that? Use the hosted version

For production workloads — bulk, RAG-ready chunks, SRT/VTT, 100+ languages, hosted residential proxies, retries, and uptime monitoring — use the Apify actor (same engine, run for you at scale, pay per transcript):

→ YouTube Transcript Scraper on Apify

Also available: YouTube Comments Scraper.

How it works

1. Fetch the watch page, parse ytInitialPlayerResponse for caption tracks.
2. Mint a PoToken bound to the video ID via BotGuard (bgutils-js).
3. Request the caption track with &pot=<token>&c=WEB&fmt=json3 and parse the segments.

License

MIT