Phone Whisper is an Android dictation app that records speech, transcribes it, and inserts the result into text fields across apps.
Phone Whisper supports two transcription modes.
In local mode, audio is processed on-device using local speech recognition models. Audio does not leave the device.
In cloud mode, recorded audio is sent directly from the device to OpenAI's transcription API to generate text.
If optional cleanup is enabled, the transcribed text is also sent directly from the device to OpenAI's chat API to improve punctuation, capitalization, and clarity.
If you use cloud features, your OpenAI API key is stored locally on your device in app storage and used to authenticate requests sent directly to OpenAI.
I do not operate a relay server for these requests.
Phone Whisper uses Android Accessibility Service only to identify the currently focused text field and insert dictated text after you explicitly interact with the floating overlay button.
Phone Whisper is not designed to monitor browsing, collect screen content for analytics, or perform background automation.
I do not run a backend for Phone Whisper and do not collect user accounts, analytics, or uploaded recordings myself.
Third-party services you choose to use, such as OpenAI, may process data according to their own terms and privacy policies.
For questions about privacy, contact: pol.avms@gmail.com