Private conversation voice notes

A voice note belongs to a shared conversation and its original sponsoring business. Current business operators who have both posting and file-sharing permission can upload one. Thread-only guests can listen to posted recordings inside their permitted message history; they cannot upload files or spend another business’s storage allowance.

Record, listen, and review

Open a shared conversation and find Private voice note. On browsers that support microphone capture and WebM/Opus or M4A recording, choose Record voice note, then stop and listen. Cancel discards the local recording. Recording stops before one minute. If recording is unavailable or microphone permission is denied, upload an MP3, WAV, M4A, or WebM file, or use the text composer.

Files must be at most 8 MB and decode to between one and 60 seconds. Duration tags and file extensions are not trusted. Processing checks the complete decoded audio and produces a metadata-stripped playback copy. An invalid or overlong recording cannot be posted; it is not silently shortened.

Write an accurate transcript and review the displayed media ingestion charge before uploading. Storage requires an activated private provider and the business owner’s enabled, capped usage allowance. The ingestion charge includes the playback copy; it is not a new charge each time someone listens. No transcription provider is activated or called. Transcript text is supplied and reviewed by you.

An upload is a private draft. Refresh its processing status, listen again, correct the transcript if needed, and choose Review voice note and access. Only Approve and post voice note publishes the recording and its frozen transcript to eligible participants. The voice-note composer selects no email notices. It never sends a text or makes a public audio link.

Current access and retained history

Playback requires a current named participant with access to that message’s history. Revoked, expired, foreign-business and earlier-history access is denied. Copying a playback URL does not grant access. The original recording and storage location are never exposed in the player or generic file pickers. Playback uses keyboard-operable browser controls and never autoplays.

Posted transcripts stay as reviewed. A later draft correction cannot rewrite them. Remove my business’s playback disables access and queues file cleanup while preserving the posted transcript, attribution and file evidence. It does not erase another participant’s memory or an audio copy they already obtained. Other currently authorized participants keep access if one participant is revoked; revocation does not delete the shared message.

Unposted drafts expire after 30 days. Explicitly removed and expired files use durable cleanup. Storage whose acceptance is uncertain remains reserved until its exact saved artifact is reconciled; a lost response does not authorize another charge or a new file identity.

Interrupted uploads and processing

If upload confirmation is missing, retry the original file and transcript before starting a different upload. After a refresh, choose Resume original upload on the interrupted draft and reselect the exact original file. This preserves its saved request, transcript and original charge, even if pricing changed. Transcript corrections become available once processing finishes. A failed processing job can be retried after its source and current permissions are checked again. Paused spending or revoked permission prevents new storage writes. Already accepted bytes can still be reconciled and removed without fabricating a successful post.

Voice notes do not provide automatic dictation, guest voice messages on public websites, music uploads, or transcription accuracy guarantees. Production storage and codec execution remain part of coordinated activation.

Dictate an editable draft

Choose Dictate text in a rich-message editor or private SMS composer, or Dictate transcript beside a voice note. Start browser dictation deliberately, stop, edit the words, then choose Insert reviewed text. The normal preview and send approval still follow. The transcript can also be typed. Browser speech support varies and may send microphone audio to the browser provider for recognition. This is microphone dictation, not automatic transcription of an uploaded recording. Your browser’s microphone permission and language selection control capture.

Record from the native app

In Shared conversations, open a thread and choose Record a voice note. Microphone access is requested only when you choose to record. Capture stops after about one minute or when you leave the foreground. Listen locally and supply an accurate transcript before approving the displayed private-storage charge. A recording you have not uploaded is cleared when you leave the conversation.

Refresh private drafts after processing, select the recording, and use the message preview to check its frozen transcript and chosen email recipients. Then approve the shared message. Uploading alone never posts a message. An uncertain upload retains the original request while this screen stays open; retry it before starting another. Recordings that reached the server also appear in the private draft library.

Playback uses the current signed-in account and participant permissions. It pauses in the background; there is no public download URL or background microphone mode. Native capture/playback, audio interruptions, permission denial and physical-device recovery remain part of the combined activation rehearsal.