A pronunciation guide only tells you what someone else thinks a name sounds like. A selfie is the person saying it themselves: a few seconds of an interview, podcast or talk, cued to the moment they say their own name. Audiobook narrators, producers and journalists can play it before going on the record.
What the labels mean
Verified by ear somebody listened and confirmed it's that person saying their own name at that second.
Best guess a clip has been found but nobody has confirmed the exact second yet. Play it, and if it's right, tap “That's them”.
Disputed more people have flagged the clip than confirmed it.
Clip needed the name is listed with a respelling from a secondary source, but there's no clip yet. The finder helps you track one down.
How to help
Anyone can add a person, add a clip, confirm one, flag one or suggest a better time. Nothing is ever deleted: every change is kept in the entry's history, so a wrong edit can always be put back.
Where the clips come from
Clips play from YouTube, Spotify and other sites where they were published; nothing is copied or re-hosted. The finder searches spoken words in YouTube captions through the YouGlish widget. Playing a clip loads YouTube's player, which sets its own cookies; see YouTube's terms and Google's privacy policy.
Silk Selfies is part of the Silk Suite of audiobook production tools.