Standalone forensic-grade voice distorter. Three modes — LIVE playthrough (mic → processing → output, including over Bluetooth SCO headsets), RECORD a processed voice memo, or IMPORT an existing audio file and anonymize it. Same DSP that powers Anon XMPP / Anon Mumble's in-call voice changer, in a standalone app. LIVE mode — AudioRecord → STFT + random-phase per FFT bin + 47 Hz ring modulation → AudioTrack, ~10 ms one-way latency on the phone's speaker. Bluetooth SCO route supported (both mic and earphone on a paired headset) RECORD mode — tap once to capture a processed voice memo, save as .wav, hand off to the Android share sheet (drop into Signal, email, your tip portal…) IMPORT mode — pick any audio file via Storage Access Framework (wav, m4a, mp3, ogg/opus, flac, …). MediaCodec decodes, the DSP runs, an anonymized .wav drops out. Two forensic modes plus OFF: ROBOT (information-theoretically irreversible — random phase per bin destroys speaker micro-structure; cannot be inverted) and MASK (envelope-flattened + random phase — sounds more human, still destroys speaker ID). No plain pitch-shift mode (pitch shift is reversible) No internet permission, no analytics, no Google Play Services, no auto-backup FLAG_SECURE blanks the app in Recent Apps and blocks in-app screenshots Free software under GPLv3 F-Droid repo: https://anonymous.gratis/fdroid/repo Tip portal: http://ieyezgeojxw73hv4szrkbreea3rd6ri7xfbevngi63uabothn226euyd.onion