# ChipChat: Low-Latency Cascaded Conversational Agent in MLX

## TL;DR

ChipChat is a low-latency cascaded conversational agent that integrates multiple streaming components on-device, enabling real-time voice interaction without sacrificing privacy or performance.

## Contribution

The paper introduces ChipChat, a novel low-latency cascaded system with architectural and streaming optimizations for real-time on-device voice agents.

## Key findings

- Achieves sub-second response latency on Mac Studio without GPUs.
- Maintains user privacy through complete on-device processing.
- Demonstrates that redesigned cascaded systems can overcome latency limitations.

## Abstract

The emergence of large language models (LLMs) has transformed spoken dialog systems, yet the optimal architecture for real-time on-device voice agents remains an open question. While end-to-end approaches promise theoretical advantages, cascaded systems (CSs) continue to outperform them in language understanding tasks, despite being constrained by sequential processing latency. In this work, we introduce ChipChat, a novel low-latency CS that overcomes traditional bottlenecks through architectural innovations and streaming optimizations. Our system integrates streaming (a) conversational speech recognition with mixture-of-experts, (b) state-action augmented LLM, (c) text-to-speech synthesis, (d) neural vocoder, and (e) speaker modeling. Implemented using MLX, ChipChat achieves sub-second response latency on a Mac Studio without dedicated GPUs, while preserving user privacy through complete on-device processing. Our work shows that strategically redesigned CSs can overcome their historical latency limitations, offering a promising path forward for practical voice-based AI agents.