Instant voice feedback
Streaming pipeline speaks the earliest tokens ASAP for a “go-now” feel.
Real-time AI Voice Coach for CS2
Ask naturally, interrupt anytime, and get round-winning calls in the moment.
Real-time STT & TTS
GSI + Smart images parsing
Adaptive model response
Google/Piper/Coqui TTS engines
Why Clutch
Grounded in your match
Combines CS2 GSI with targeted screenshots (radar + alive counts) for context-aware calls & local timer badge.
Voice that keeps up
Uses Google Chirp 3 HD when available, with automatic Piper/Coqui fallback chosen by VRAM.
Key features
- Windows desktop UI
- Auto GSI setup
- Bomb & round timers
- “Hey Jarvis” wake word
- Realtime STT with partials
- CS2 GSI + multi-region screenshots
- LLM streaming brain
- Instant TTS: Google / Piper / Coqui
- Immediate speech stop on interrupt
How it works
-
1
Wake
Say “Hey Jarvis”. Wakeword + VAD starts listening.
-
2
Hear
Realtime STT streams partials so you see what’s heard instantly.
-
3
See
GSI + targeted screenshots (radar & alive counters) ground the plan.
-
4
Think
The brain reasons in context for concise, instantly usable calls.
-
5
Speak
Streaming TTS talks as tokens arrive. Interrupt anytime to re-ask.
Get Clutch
Please head to GitHub for the installation manual and demo.