subtitles

The subtitles skill provides AI agents with direct, efficient access to YouTube transcript data via the TranscriptAPI. This skill is essential for users who need to process spoken content from video platforms without relying on heavy browser automation or unstable scraping tools. It is designed for students, researchers, content creators, and developers who need programmatic access to what is said in a video, whether for generating summaries, creating study aids, or enabling accessibility features for the deaf and hard of hearing.

• Retrieves accurate, timestamped transcript segments for any public YouTube video using a simple API endpoint.
• Supports both raw text output (ideal for translation and summarization tasks) and structured JSON (ideal for sync-reading or captioning applications).
• Allows for optional metadata inclusion, such as video title, channel name, and description.
• Works seamlessly with major AI agent frameworks, requiring only a standard API key for authentication.
• Handles multi-language auto-generated subtitles, enabling users to follow foreign-language content with ease.

Usage notes and technical details:

• Requires a valid TRANSCRIPT_API_KEY. The skill includes automated flows to assist users in registering or configuring their credentials upon first run.
• Proper configuration of the User-Agent header is strictly enforced to prevent Cloudflare 403 blocks and ensure reliable connectivity.
• The tool is strictly for fetching subtitles and transcripts; it does not support account management, video uploading, or administrative YouTube functions.
• The API operates on a credit-based system, with a generous free tier available for personal or development use.
• Use the timestamped format if you are building applications that require real-time playback synchronization or highlighting specific spoken moments.
• For bulk processing, users can request transcripts for entire channels or playlists when combined with related YouTube skills, making it a powerful tool for deep-dive research into video databases.