Get voice memos
Returns a keyset-paginated page of the user's voice memos (MessageV6) ordered by last_updated_at, optionally narrowed to a workspace or folder (folder_id=null for the ones in no folder). Deleted voice memos come back as tombstones unless include_deleted=false. Because the feed is keyed on last_updated_at, a direction=newer sync from a stored cursor returns memos created since it and older memos changed since it (deleted, moved to or out of a folder, edited), so the same memo can come round again; de-duplicate by id and keep the latest. Successor to GET /v3/messages/voicememo and GET /v4/messages/voicememo, whose date filter is also on last_updated_at. Paging semantics are otherwise identical to GET /v6/messages. MessageV6 differs from v5 in three ways: thread_id (the parent id, or the message id itself when it is not a reply) replaces parent_message_id; the audio rendition and its transcript, AI summary, language flags and time codes are nested under content, with the waveform as a compact base-36 string; and the requesting user's own state (heard_ms, label_ids, notify, notes, last_heard_update) is included where the read path carries it.
Request
- Base URL: https://api.carbonvoice.app
- URL: https://api.carbonvoice.app/v6/voicememos
- Auth: one of:
- API key in header pxtoken
- OAuth 2 (implicit, authorization URL https://api.carbonvoice.app/oauth/authorize)
Query parameters
Anchor date to page from (older/newer than this instant). Required for the first page; omit it on subsequent pages and pass cursor instead. Ignored when cursor is provided.
Opaque pagination cursor returned as next_cursor from a previous response. When provided, it supersedes date. A cursor issued while has_more was true continues strictly from the last message of the previous page (no overlap). The cursor a final direction=newer page issues marks the newest message seen: sent back later it returns what was created or changed since, and may include up to 4s of already-seen messages behind that boundary while it is recent, so a message whose write landed late is not lost; de-duplicate by id. A quiet feed converges to an empty page. Keep direction the same across a cursor sequence.
Maximum number of messages to return (default: 200).
Language for message content (e.g. transcript, audio). Original language used if not provided.
Include presigned URLs for audio files (default: false).
Only voice memos in this workspace.
Only voice memos in this folder. Send the literal null for the voice memos in no folder; omit it for all of them.
Include deleted voice memos as tombstones (status: deleted) so a syncing client can drop them (default: true).
Response
Changes
No recorded changes to this operation across all 1 revision of this API.