Official documentation
VocalCode privacy notice
VocalCode, 2026
Read the originalRead on October 22, 2026
- What it is
- The maker's privacy notice. It says the current free desktop release recognizes speech locally and does not send audio or transcripts to the maker's servers, and it describes the local dictionary.
- How much weight it carries
- The maker's own notice, read 2026-10-22. The notice describes the current free release and also an earlier paid release; the quotes used here are about the current release.
Facts taken from this source
- VocalCode's privacy notice says dictation audio is processed in memory for recognition and is not intentionally written to disk. The notice gives no period.
“Dictation audio is processed in memory for recognition and is not intentionally written to disk.” (Privacy notice, local dictation paragraph)
- VocalCode's privacy notice says that when the user starts a meeting, audio is written as crash-recovery chunks, and that those chunks are deleted after successful transcription and note generation unless the user selects Keep audio.
“selected microphone and/or system audio is instead written as bounded local crash-recovery chunks while recording. Those chunks are deleted after successful transcription and note generation unless you select Keep audio” (Privacy notice, meetings paragraph)
- VocalCode's privacy notice says meetings are intentionally stored on disk until the user deletes them under the retention setting.
“meetings are intentionally stored on disk until deleted under the retention setting.” (Privacy notice, short version)