




I built a Stream Deck plugin that auto-ducks your music/game the instant you talk — SpeechDuck
You're mid-song, someone drops into your Discord call, and now you're scrambling for a mixer, an OBS audio filter, or a second monitor just to turn the music down before they can hear you — and turn it back up two minutes later, if you remember. I got tired of doing this manually every single session, so I built SpeechDuck: a Stream Deck plugin that does real sidechain ducking, the same trick broadcast radio and podcast studios have used for decades, entirely from one key.
How it works: the moment your mic picks up your voice — or a conferencing app starts talking back — SpeechDuck automatically pulls your music/game/desktop audio down to whatever level you set, and brings it back up the instant you go quiet. One key arms/disarms the whole thing, and while it's running that same key doubles as a live level meter so you can see it's actually listening.
I use it every day just for my own listening, not only on stream — but this is genuinely one of those features that's way bigger for streaming than for solo use. Your viewers stop noticing your audio entirely, which is exactly the point: no more chat complaining that the music is drowning out you or your friends.
https://reddit.com/link/1vsnhka/video/i6fuoc436ckh1/player
What you can configure
- Triggers — any microphone, any output device, or a specific running app (Discord, a game, anything) — mix as many as you like. An app you add before it's even running gets picked up automatically the moment it launches.
- Targets — whatever gets turned down: music apps, games, or an entire output device, each with its own "volume while ducked" level.
- Attack / Sustain / Release — how fast it ducks, how long it stays down through pauses in speech, how gently it comes back up.
- Drag-to-set threshold directly on the live level meter — no guessing at numbers.
- Live meters everywhere — every trigger and target shows real signal level, in settings and right on the key.
- ~60Hz polling, so the reaction to your voice is effectively instant — no audible lag between the first syllable and the music dropping out of the way.
- 9 interface languages built into the settings panel (EN/DE/FR/ES/JA/KO/ZH-CN/ZH-TW/RU).
Get it
€9.99, one-time purchase, no subscription: https://marketplace.elgato.com/product/speechduck-ea695e8d-fe4e-4581-9319-ba1df369e49c
💻 GitHub (support, bug reports, feature requests): https://github.com/SquierStream/streamdeck-speechduck
Happy to answer questions about how the ducking/envelope math works, or take feature requests — this is actively maintained.
My other Stream Deck plugins, if you're into this kind of thing:
- 🎵 Now Playing for Dials — turns your Stream Deck+ touchscreen into a full-width media panel (album art, title, progress, shuffle/repeat) with volume on the dials: https://marketplace.elgato.com/product/now-playing-for-dials-cade76d9-d5ed-4f1d-a8d9-9a249cdf6a93
- 📊One Key System Monitor — five readings on a single key: CPU load and temperature, GPU load and temperature, and memory in use — as numbers, as bars, and as colour that shifts from green to amber to red while things heat up.: https://marketplace.elgato.com/product/one-key-system-monitor-be45d32e-38b8-4384-ad67-9d95e5f4ec59
One question for you all: would a free Demo version be worth putting together — 1 trigger, 1 target, English-only interface (so no multi-app setups or other languages, but honestly enough for the common "one mic ducks one thing" case)? Let me know if that's something people actually want before I put the work into shipping it.