Repository navigation
SDR-Kit 3.6.0
Added
- Voice descrambler (Radio → Scrambled voice): undoes the frequency-inversion "scrambler" / "voice privacy" of many
cheap walkie-talkies. Pick the inversion point (2600–3500 Hz, 2632 and 3300 are common); recordings and the scanner get the
descrambled voice. Covered bytests/test_voice_inverter.py. It does not defeat digital encryption or rolling scramblers. - RTL-SDR Blog V3 support is now tested in CI:
tests/test_rtl_models.pyruns the panel's driver code against a fake
librtlsdr that reports a V3 (R820T) or a V4 (R828D) and checks detection, gains, direct sampling below 24 MHz on the V3
only, sample delivery and the advanced receiver's HF profiles. The README says the V3 works. - README (EN/ES): "Monitor your business from anywhere" use case — leave the kit at a business that runs on
walkie-talkies and listen, record and check the activity log from abroad over Tailscale (needs internet at the business).
Install: curl -fsSL https://raw.githubusercontent.com/romartak/sdrkit/main/get.sh | sudo bash · Update: sudo sdrkit update