Local-first audio workspace
Turn long audio into searchable memory.
Import locally, transcribe on your own mere.run node, and search every word, without shipping your audio to a hosted service.
mere.media is invite-only right now. Join the waitlist on mere.world and we will open workspace access from there.
- 12:18 Speaker C
The action item is to tag source confidence, speaker confidence, and whether the quote can be reused.
- 18:14 Speaker B
Search should bring back the policy context even when the user only remembers the phrase cooperative reporting.
- 25:19 Speaker A
Keep lexical search visible because embeddings can lag behind local transcription on a long recording.
Library
Long audio, made scannable.
Talks, calls, interviews, audiobooks, and field recordings settle into one workspace with transcript status, segment counts, import history, and source-linked search.
Workflow
field-recording.m4a
localwhisper-large-v3
running128 segments
readyjump to the second it was said
privatemere-media import file ./recording.m4a The CLI owns local processing. The browser is the ledger and search surface.
Private by default
Your transcripts are not a public corpus.
Audio is transcribed on your own mere.run node. The workspace syncs transcripts, segments, and search state; the media itself never has to leave your machine.
Join waitlist on mere.world