---
title: "Pace vs RCLI"
description: "An evidence-based comparison of Pace and RCLI: tier picker + trust surfaces vs MetalRT speed + RAG."
canonical: https://heypace.app/compared/rcli/
last-updated: 2026-08-27
---
# Pace vs RCLI

An evidence-based comparison of Pace and RCLI: tier picker + trust surfaces vs MetalRT speed + RAG.

## Product posture

- **Product:** [RCLI](https://github.com/RunanywhereAI/RCLI)
- **Maintainer:** RunanywhereAI
- **License:** MIT; open source under MIT
- **Runtime posture:** on-device
- **Speech-to-text:** Local — Zipformer (streaming), Whisper base.en, Parakeet TDT 0.6B
- **Reasoner:** Local — Qwen3 0.6B/4B, Llama 3.2 3B, LFM2.5 1.2B (MetalRT or llama.cpp)
- **Text-to-speech:** Local — Piper, KittenTTS, Matcha, Kokoro (28 voices)
- **Screen-aware:** Yes

## What the competing product does well

- Sub-200ms end-to-end voice latency via proprietary MetalRT GPU engine (up to 550 tok/s).
- Local RAG over documents with hybrid vector + BM25 retrieval.
- 20+ models across LLM, STT, TTS, VLM, VAD, and embeddings — all on-device.

## Where Pace differs

RCLI's MetalRT engine hits sub-200ms latency and 550 tok/s — a serious performance benchmark Pace should measure against. RCLI has local RAG Pace doesn't ship. Pace has the tier picker, trust surfaces, and MCP integration RCLI lacks. 1,523 stars — the most-trafficked on-device competitor.

This comparison does not claim ratings, market rank, or user-review scores.