Bring voice input to your game with fully offline, real-time speech-to-text. No internet required, no API keys, no subscriptions. Everything runs locally on the player's machine.

Auvyra STT is powered by Whisper and SenseVoice via ONNX Runtime, delivering accurate transcription directly inside Unreal Engine. Choose Whisper for maximum accuracy across 99 languages, or SenseVoice for up to 3× faster transcription with automatic language detection.

The plugin exposes a clean, easy-to-use Blueprint API. Create a session with one node, bind the OnSpoken event, and start receiving transcribed text instantly.

📖 Documentation

🎮 Demo Build

Features:

• 🔒 100% Offline. No internet required. No API keys. No subscriptions. No usage limits. Runs entirely on the player's machine.

• 🧠 Two STT Models. Whisper for 99-language accuracy; SenseVoice for ~3× faster transcription with auto language detection.

• 🌍 99 Languages. Full multilingual support via Whisper. SenseVoice supports Auto, English, Chinese, Japanese, Korean, and Cantonese.

• 🧩 Full Blueprint API. Every feature is exposed to Blueprint. No C++ required.

• 🎙 Automatic Voice Activity Detection. Built-in VAD detects speech start and end automatically. Fully tunable silence timeout, max and min speech duration.

• 📁 Audio File Transcription. Transcribe WAV, MP3, and FLAC files at runtime with a single Blueprint node. Result delivered via OnSpoken.

• 🔊 Microphone Selection. List all available audio input devices and switch between them at runtime without restarting the session.

• ⚡ Runtime Model & Language Switching. Change the STT model or language at any time without recreating the session.

• 🧵 Zero Game Thread Impact. All transcription runs on a dedicated background thread. Your frame rate is never affected.

• ⏹ Cancel Utterance. Discard the current in-progress speech segment without stopping the microphone.

• 📦 Ready-to-Ship Builds. Works in both Development and Shipping builds with no extra setup.

Whether you're building voice commands, NPC interaction, accessibility input, in-game dictation, or any system that needs speech, Auvyra STT handles it without a server, without a subscription, and without blocking your game thread. All transcription runs on a dedicated background thread.