ytranscript

Fetch YouTube video transcripts via Innertube API with AI-powered features. Zero YouTube API key required.

Quick Start

# Get transcript (no install needed!)
npx ytranscript VIDEO_ID

# Summarize a video
npx ytranscript VIDEO_ID --summarize

# Ask questions about video content
npx ytranscript VIDEO_ID --ask "What are the main points?"

# Translate transcript
npx ytranscript VIDEO_ID --translate Spanish

Installation

npm install -g ytranscript

CLI Usage

Basic Transcript

# Text with timestamps (default)
ytranscript VIDEO_ID

# JSON format
ytranscript VIDEO_ID --format json

# SRT subtitles
ytranscript VIDEO_ID --format srt -o subtitles.srt

# Plain text without timestamps
ytranscript VIDEO_ID --no-timestamps

# List available languages
ytranscript VIDEO_ID --list-languages

# Fetch specific language
ytranscript VIDEO_ID --lang es

AI Features

Requires OPENROUTER_API_KEY environment variable (get one at openrouter.ai).

# Summarize the video
ytranscript VIDEO_ID --summarize

# Ask a question about the video
ytranscript VIDEO_ID --ask "What is the main argument?"

# Translate to another language
ytranscript VIDEO_ID --translate Japanese

# Use a specific model
ytranscript VIDEO_ID --summarize --model sonnet
ytranscript VIDEO_ID --ask "Explain the key concepts" --model gpt-4o

Available Models

Shortcut	Model	Best For
`haiku`	Claude 3 Haiku	Fast, cheap (default)
`sonnet`	Claude 3.5 Sonnet	Balanced
`opus`	Claude 3 Opus	Best quality
`gpt-3.5`	GPT-3.5 Turbo	Fast, cheap
`gpt-4o-mini`	GPT-4o Mini	Balanced
`gpt-4o`	GPT-4o	Best quality
`gemini-flash`	Gemini Flash 1.5	Fast
`gemini-pro`	Gemini Pro 1.5	Best quality

Or use any OpenRouter model ID directly.

Output Formats

Format	Description
`text`	Plain text with timestamps (default)
`json`	Structured JSON with start, duration, text
`srt`	SubRip subtitle format

Programmatic Usage

import { getTranscript, listLanguages, formatText } from "ytranscript";
import { summarize, askQuestion, translate } from "ytranscript/ai";

// Get transcript
const transcript = await getTranscript("VIDEO_ID");
for (const segment of transcript) {
  console.log(`[${segment.start.toFixed(1)}s] ${segment.text}`);
}

// AI features (requires OPENROUTER_API_KEY)
const text = formatText(transcript, false);
const summary = await summarize(text, { model: "anthropic/claude-3-haiku" });
const answer = await askQuestion(text, "What is this about?");
const japanese = await translate(text, "Japanese");

Types

interface TranscriptSegment {
  start: number;    // Start time in seconds
  duration: number; // Duration in seconds
  text: string;     // Transcript text
}

interface LanguageInfo {
  code: string;     // Language code (e.g., "en", "es")
  name: string;     // Language name
  isAuto: boolean;  // Whether auto-generated
}

How It Works

Transcript: POST to YouTube Innertube API with Android client impersonation
AI Features: Send transcript to OpenRouter API for processing

No YouTube API key needed. AI features require OpenRouter API key.

Requirements

Node.js 18+ (uses native fetch)
OpenRouter API key (for AI features only)

Limitations

May be blocked on datacenter IPs (works fine on residential)
Age-restricted videos may not work
YouTube may change their internal API

License

MIT