Mayari icon

Mayari

Read PDFs, EPUBs and Word docs aloud with local neural voices

Free Text to Speech Open Source
Listing checked 8 Aug 2026

Overview

Mayari is a local-first PDF reader for macOS that combines document reading with high-quality neural text-to-speech. Powered by KokoroSwift running on Apple's MLX framework, Mayari generates natural-sounding audio entirely on-device with zero external servers. Read PDFs, DOCX, and EPUB files while listening with eight British voices at adjustable speeds from 0.5x to 2.0x. Features include quote capture with page citations, highlight mode, background audiobook generation with a persistent queue-based job system, markdown export, and a built-in text editor with TTS support. The clean, distraction-free interface follows Apple's Human Interface Guidelines with dark mode support, a resizable 60/40 split layout, and smooth PDF navigation with zoom controls.

Pricing: Free

Minimum macOS: 15.0 (Sequoia)

Architecture: Apple Silicon

Key Features

  • 54 Kokoro voices across 9 language catalogs
  • Reads PDF, DOCX and EPUB aloud
  • Background audiobook queue with retry and cancel
  • Audio generation 3-5x faster than real time
  • Quote capture with page numbers (Cmd+D)
  • Highlight mode (Cmd+H)
  • Markdown export with formatted citations
  • Fully native - no Python, servers or external dependencies
  • KokoroSwift running on Apple MLX
  • Offline after the initial model download

Tags

voice synthesis