Back to Portfolio

Local Talk Pipeline

Ultra-low latency, locally hosted speech-to-text translation pipeline for streaming.

Classification AI & Edge Computing
System Status local_host
Infrastructure
Docker FastAPI Python Silero VAD Whisper

Technical Case Study

An ultra-low latency, locally hosted speech-to-text and translation pipeline. It captures local audio, detects voice activity using Silero VAD, and transcribes it using Faster-Whisper. Features hardware acceleration via NVIDIA GPUs over Docker.