Consent and privacy
Last modified: 25 July 2026
What happens to audio
- Audio is analysed in memory. It is never written to disk by the plugin.
- The only destination for an utterance is the processor you configured — your own container, or the Sniffer Studio cloud if you subscribe. Nothing else receives it.
- There is no always-on recording, in any mode.
- What a flag produces is text: the player, the severity and the transcript line, delivered to staff.
If you self-host the processor, audio never leaves your infrastructure at all.
What players are told
Every player is shown a disclosure before their microphone is read. The wording is yours:
consent:
message: Voice chat may be transcribed for moderation. Audio is not stored.Keep it accurate and keep it in the language your players speak. It is your disclosure to your community, not ours — say that voice is analysed for moderation, and do not say "recorded" if you are not recording.
Opting out
consent:
opt-out-mode: disable-voice| Mode | Effect |
|---|---|
disable-voice (default) | A player who refuses transcription loses voice chat. They can still play and still use text chat. |
honor | They keep talking, unmoderated. |
disable-voice is the default on purpose: moderation you can decline is not
moderation. honor exists because some communities genuinely want it, but on a public
server it hands anyone who wants to be abusive an opt-out button.
Who is responsible for what
You are the data controller for your players. VoiceSniffer is compliance tooling, not a legal guarantee — the disclosure mechanism, the opt-out behaviour and the no-storage default are ours; the decision that your use of them is lawful where you operate is yours.
Sniffer Studio's own site privacy notice is at snifferstudio.net/privacy.