Community Tutorial : How to do on-device offline speech to text in Unreal Engine without any cloud API key.

While working on my game, I needed a reliable Speech-to-Text (STT) solution inside Unreal Engine.

Instead of relying on a cloud based service, I started building a solution around whisper.cpp. What began as a feature specifically for my game gradually turned into something I realized could be useful for other Unreal Engine developers as well.

So I decided, why not turn it into a proper plugin and release it on Fab

:studio_microphone: Introducing WhisperSTT - Offline Speech-to-Text for Unreal Engine

Built on whisper.cpp, WhisperSTT brings private, on-device speech recognition directly into Unreal Engine with no cloud APIs, no API keys, and no per request transcription costs.

Designed with game developers in mind, WhisperSTT makes it easier to build voice driven gameplay, NPC dialogue systems, voice commands, accessibility features, and interactive experiences.

:sparkles: Key Features:
:microphone: Offline, on-device Speech-to-Text
:video_game: Blueprint-first integration
:speaking_head: Push-to-Talk & Continuous Listening
:speaker_high_volume: Built-in Voice Activity Detection (VAD)
:high_voltage: Fully asynchronous transcription
:globe_showing_europe_africa: Multilingual speech recognition
:bullseye: Fuzzy speech matching for dialogue choices & voice commands
:locked: No cloud dependency
:laptop: Built on whisper.cpp

Whether you’re creating voice-controlled gameplay, experimenting with speech recognition in Unreal Engine, or building more immersive ways for players to interact with your game, WhisperSTT provides the foundation to bring voice into your Unreal Engine projects.

:rocket: WhisperSTT is now available on Fab!

FAB Link