Add Transcripts

[ephes]

• Add a transcript model with vtt format for feed and json for the podlove player
• Add the transcript tab to the podlove player

Should we add a new transcript model or just reference the file in the json field of an audio model? A separate transcript model is probably cleaner but it's more complicated to avoid additional queries.