mr-transcript 0.1.2

A simple wrapper for youtube-transcript-api to get YouTube video transcripts with ease.

Mr. Transcript

A convenient wrapper for the youtube-transcript-api library designed to retrieve YouTube video transcripts quickly and reliably.

Key Features

• Automatic URL Handling: Supports various YouTube link formats, including youtube.com, youtu.be, shorts, and embed.
• Intelligent Search: The package first looks for manually created transcripts; if they are unavailable, it automatically switches to YouTube's auto-generated ones.
• Timecodes: Option to add timestamps to each text block.
• Language List: Quickly retrieve a dictionary of all available languages for a specific video.
• Type Safety: Full support for type annotations for better development experience.

Installation

Install the package via pip:

pip install mr-transcript

Or using uv:

uv add mr-transcript

Quick Start

from mr_transcript import get_transcript, get_languages

url = "https://www.youtube.com/watch?v=dQw4w9WgXcQ"

# 1. Check available languages
languages = get_languages(url)
print(f"Available languages: {languages}")

# 2. Retrieve transcript text (e.g., in English)
if "en" in languages:
    text = get_transcript(url, language="en", timecodes=True)
    print(text[:500])  # Prints the first 500 characters

AI Agent Skill

This repository includes a specialized skill for AI agents (like Gemini CLI). It helps the agent provide expert assistance with Python coding, URL parsing, and integrating this library into your projects.

To install the skill, run:

npx skills add BogdanovychA/mr-transcript --skill mr-transcript

Function Descriptions

get_transcript(video_id, language, timecodes=False)

Retrieves the full transcript text.

• video_id: Video ID or a full link.
• language: Language code (e.g., 'en', 'uk').
• timecodes: If True, adds the start time to each text block.

get_languages(video_id)

Returns a dictionary of all available languages for the video.

• Format: {"language_code": "language_name"} (e.g., {"en": "English"}).

License

This project is licensed under the MIT License.