HaykTarkhanyan/voice-interview-armenian
A voice interview agent for the Armenian language built on LiveKit Agents, implementing a cascading STT -> LLM -> TTS pipeline. It uses ElevenLabs for Armenian speech-to-text and text-to-speech, Google Gemini as the LLM, and Silero for voice activity detection. The project includes research documentation on the low-resource Armenian language stack and provides both local console and LiveKit server deployment options.
An Armenian-language voice assistant pipeline that handles speech-to-text, intent matching against an FAQ, knowledge-...
ArmSpeechTT is a fine-tuned Whisper model for Armenian speech-to-text, trained on Common Voice data with a webcam dem...
An end-to-end voice AI customer support agent for Armenian banks that understands and speaks Eastern Armenian. It use...
ArmTTS-Cpp is a C++ library for Armenian text-to-speech synthesis that acts as a client wrapper for a RapidAPI-hosted...
A speech-to-text model for Armenian language using Wav2Vec2-BERT, fine-tuned on the Armenian Common Voice dataset. Th...
A personal playground repository for testing and comparing Armenian speech recognition models. Contains Jupyter noteb...