Dictation and transcription
Whisper privacy rating
Open-source speech recognition model and Python command-line tool from OpenAI that transcribes and translates audio in many languages on the local computer.
Summary
Whisper scores 87 out of 100 (grade B) on the dictation and transcription criteria. It meets 5 of 6 criteria: open source, no trackers or telemetry, no ads or data sales, runs on the device and no training on recordings. It does not meet independent audit. It is based in the United States: Five Eyes member; subject to the US CLOUD Act.
Score 87 out of 100. How scoring works
Criteria
-
Yes
Open source Weight 3 of 3
Is all the source code needed to run the product public?
Code and model weights are MIT-licensed.
-
Yes
No trackers or telemetry Weight 3 of 3
Are the website and apps free of third-party trackers, with any analytics anonymous and any telemetry off by default?
No telemetry or analytics in the source code.
-
Yes
No ads or data sales Weight 2 of 3
Is the product funded without advertising, ad targeting or selling user data?
Free open-source software with no ads.
-
No
Independent audit Weight 2 of 3
Has an independent security or privacy audit been published within the last three years?
No independent audit is published.
-
Yes
Runs on the device Weight 3 of 3
Is speech converted to text on the device, without sending audio to a server?
The model runs on the local computer; audio is not sent to a server.
-
Yes
No training on recordings Weight 2 of 3
Are recordings and transcripts kept out of model training by default?
Processing is entirely local.