Implement the podcast:transcript tag with multiple supported formats

Verified steps

1. Generate transcript files in at least two formats for broad compatibility: WebVTT ('text/vtt') for in-player synchronized display and SRT ('application/x-subrip') or plain text ('text/plain') for accessibility tools and screen readers
2. Add a separate 'podcast:transcript' tag inside the RSS 'item' for each available transcript format and language — multiple tags are explicitly supported and recommended: '<podcast:transcript url="..." type="text/vtt" language="en" />'
3. Include the 'language' attribute on each transcript tag using a BCP-47 language code (e.g., 'en', 'fr', 'de') — this allows apps to select the appropriate transcript for the listener's device language setting
4. For JSON transcripts (MIME type 'application/json'), follow the Podcastindex JSON transcript spec which uses a 'segments' array with 'speaker', 'startTime', 'endTime', and 'body' fields per utterance — this format enables speaker identification that VTT and SRT do not support
5. Host transcript files at stable, permanent URLs — the URL is embedded in every copy of the RSS feed and cannot be changed retroactively in feeds already cached by directories
6. Verify that each transcript file is accessible without authentication and that the server returns the correct Content-Type header matching the declared MIME type in the 'type' attribute