Vois

Introduction to Vois

Vois is a desktop application for text-to-speech (TTS) synthesis that operates entirely offline. Designed as an integrated voice production studio, it enables users to script, generate, arrange, master, and export speech audio without relying on cloud services. The software targets content creators—including podcasters, audiobook authors, YouTube narrators, documentary producers, e-learning developers, and game designers—who require high-quality, expressive, and privacy-respecting voice generation.

Unlike cloud-based TTS platforms, Vois processes all audio generation locally on the user’s machine. This eliminates per-character billing, usage caps, and the need to upload sensitive scripts to third-party servers. Built with Rust for performance and safety, Vois supports real-time processing at up to 6× speed on Apple Silicon hardware.

Key Takeaways

• Runs 100% locally: No data leaves the user’s device; all processing—including voice cloning and audio generation—occurs offline
• 63+ expressive voices across 15 semantic categories (e.g., Narrators, Villains, Machines, NPCs) and 23 languages
• Voice cloning capability: Create custom voices from 5–60 second audio samples, with all processing remaining local
• Three TTS engines: Fast, Expressive, and Multilingual—each optimized for different use cases and language support
• Integrated multi-track timeline editor with crossfades, ambient layering, and speaker-tagged dialogue support
• Professional mastering tools including LUFS normalization, de-essing, EQ, and limiter, with export presets for Spotify, YouTube, Apple Podcasts, and ACX
• Supports import from PDF, EPUB, DOCX, and web articles; includes Markdown-compatible script editor and pronunciation dictionary
• Export formats include WAV, MP3, and FLAC with configurable bitrates and sample rates

How Vois Works

Vois follows a linear yet flexible four-stage workflow: writing, casting, generating/arranging, and mastering/exporting. Users begin by writing or importing a script—either directly in the app, from a document, or via URL—and tagging speakers for multi-character dialogue. Next, they assign voices from the built-in library or cloned models to each speaker or section. Audio generation is performed locally and unlimited; generated clips appear as editable segments on a multi-track timeline where users can adjust timing, add transitions, insert ambient layers, and manage speaker order.

The final stage involves applying mastering effects—such as loudness normalization, frequency balancing, and dynamic range control—before exporting to target platforms. Export presets automatically configure settings for industry standards (e.g., ACX’s -18 LUFS requirement). All project files, audio assets, and cloned voices are stored locally and remain fully accessible even after subscription cancellation.

Core Benefits and Applications

Vois replaces multiple disjointed tools—including cloud TTS APIs, digital audio workstations (DAWs), and mastering plugins—with a single, unified application. For podcasters, it enables solo production of multi-voice episodes with intros, outros, and guest impersonations. Audiobook authors can convert manuscripts into professionally mastered, platform-compliant audio in hours rather than days. YouTube creators benefit from consistent, scalable narration for faceless or tutorial content. Documentary makers use multilingual narration and character-specific voices to localize content across regions. Game developers generate NPC dialogue in bulk while preserving stylistic consistency and linguistic accuracy. Educational content creators produce training modules and meditation guides with tailored vocal tone and pacing—all while maintaining full data sovereignty and avoiding recurring per-use costs.