Captions

**Quick judgement:** A focused and generally robust skill for pulling YouTube captions via TranscriptAPI. Excellent for accessibility, content review, and quoting, with good timestamp support. Main drawbacks are reliance on a third-party API key, credit limits, and YouTube-only scope. **Strengths** - High-quality fit for: deaf/HoH accessibility, text review of videos, and extracting timestamped quotes. - Flexible output: structured JSON with `start`/`duration` for each caption segment, or cleaner text output. - Can include metadata (title, channel, thumbnail), which helps for organizing or referencing content. - Clear instructions for API setup and for handling common API error codes. **Key risks / limitations** - **Third-party dependency:** Requires a TranscriptAPI account and API key; downtime or API changes are outside the skill’s control. - **Credits and billing:** Only 100 free credits; 1 credit per request. Heavy or automated use could exhaust credits and incur cost (HTTP 402 on exhaustion). - **Coverage limits:** Fails when a video has no captions/CC (HTTP 404) or when TranscriptAPI times out (HTTP 408). Some YouTube videos will simply not be transcribable via this route. - **YouTube-only:** Not suitable for non-YouTube sources or locally provided media files. **Recommended scenarios** Use this skill when a user: - Shares a **YouTube URL or ID** and asks for captions, transcript, closed captions/CC, or accessibility text. - Wants to **read what was said** instead of watching (e.g., long interviews, lectures, podcasts on YouTube). - Needs **timestamped segments** for quoting, clipping, or syncing with other tools. Avoid or be cautious when users: - Need transcripts from **non-YouTube** content, or - Have unclear consent or privacy concerns about sending video URLs/content to a third-party caption service.