YouTube-to-Text SDK

A minimal client for the hosted YouTube-to-Text transcription API.

⚠️ Package currently in development. In the meantime, access the API at youtubetotext.com/api

Features

• Simple, lightweight wrapper written in Python
• Supports fetching transcripts with language fallbacks
• Handles HTTP errors and returns JSON responses

Installation

pip install youtubetotext-sdk

Quickstart

from youtubetotext.client import YouTubeToText

client = YouTubeToText(api_key="YOUR_API_KEY")
response = client.transcript("VIDEO_ID", lang="es,fr")
for segment in response["segments"]:
    print(segment["text"])

API Reference

YouTubeToText.transcript(video_id: str, lang: str = None) -> dict

• video_id: YouTube video identifier (e.g. 4t2714aREe0)
• lang: Optional comma‑separated ISO language codes (en, ko,ja, etc.)
• Returns: A dictionary with the following keys:
  • source (str): Which backend provided the transcript (youtube-transcript-api, yt-dlp, whisper-cpp)
  • language (str): Human‑readable language (e.g. "English")
  • language_code (str): ISO code (e.g. en)
  • is_generated (bool): Whether the transcript was auto‑generated
  • segments (List[dict]): List of segment objects:
    • start (float): Start time in seconds
    • duration (float): Duration in seconds
    • text (str): Transcript text

Full API docs at: https://youtubetotext.com/

Configuration

Set your API key via the YTT_API_KEY environment variable:

export YTT_API_KEY="your_api_key"

Examples

See examples/python_example.py and examples/js_example.js in this repo.

Contributing

Bug reports and pull requests are welcome at https://github.com/in-c0/youtubetotext-sdk.

License

This project is licensed under the MIT License. See the LICENSE file for details.