v56

latestOpenAPI 3.1.0raw.githubusercontent.com2026-08-05311409871.8 KB
v1
onboarding
onboarding
brain-dump

Discard Brain Dump

Drop a half-uploaded take's server buffer when the user re-records.

The row and any stored audio are left alone — only the disposable part buffer is cleared.

recording_id says which take to drop. Without it the row's current take is assumed, which is wrong as soon as a second tab has moved the row on: the caller would clear a buffer that is still being filled. Callers know their own id, so they send it.

delete/api/onboarding/brain-dump

Query parameters

recording_idstring nullable

Response

Successful Response

status'recording_uploaded' | 'transcribing' | 'transcribed' | 'extracting' | 'completed' | 'failed'
input_mode'voice' | 'typed' | 'skipped'
error_codestring nullable
has_audioboolean
greeting_readyboolean