I’m going to try to give warpee ability to know who is speaking when he accesses our transcripts from gmfarcaster Not sure how the result will be but I think I can carve out a small test to prototype it Requires: - updating transcript from speaker 1, speaker 2 m, etc to actual names. Will create a front end UI for editing speaker labels but will still require someone to watch the video and map people to speaker labels - enriching transcript snippets received from embeddings which has no speaker data with speaker data from the full transcript and passing that in as enriched context