AI Tool

Voicebox Review

Name: Voicebox
Availability: OnlineOnly
Author: Stork.AI

Voicebox is a local-first, open-source AI voice studio that offers voice cloning, text-to-speech, system-wide dictation, and AI agent integration.

shipped Jun 17, 2026freemium

Domain rating30Monthly visits14K/moAI-readableblocked

Voicebox - AI tool for voicebox. Professional illustration showing core functionality and features.

Why it matters

1Voicebox is an open-source, local-first AI voice studio, initially released on February 4, 2026.

2It supports voice cloning from as little as 3 seconds of audio and offers text-to-speech generation across seven distinct TTS engines.

3The platform provides system-wide dictation into any application and integrates with AI agents via a local REST API.

4Voicebox operates without per-token fees or API rate limits, as all processing occurs on the user's local machine.

Stork’s verdict on Voicebox

Voicebox is a local-first voice studio with extensive features, but you'll need local GPU power for optimal use.

Voicebox reviewed by Stork AI · stork.ai/en/voicebox

Specs

API Docs

View Documentation →

GitHub

View Repository →

API Available

Yes, public API

overview

What is Voicebox?

Voicebox is an AI voice studio tool developed by the Voicebox open-source project that enables developers, content creators, and accessibility developers to perform voice cloning, text-to-speech generation, and system-wide dictation. It runs entirely on a user's local machine, emphasizing privacy and offering a free alternative to cloud-based solutions. This open-source application is distinct from Meta's Voicebox, a generative AI model for speech that Meta has not made publicly available. The Voicebox at voicebox.sh provides a comprehensive voice I/O stack, including a multi-track timeline editor for audio production and integration capabilities for AI agents. It supports speech generation in 23 languages and transcription in 99 languages via OpenAI Whisper.

features

Key Features of Voicebox

Voicebox provides a comprehensive suite of tools for voice manipulation and generation, designed for local execution and developer integration. Its feature set includes advanced voice cloning, diverse text-to-speech options, and robust audio production capabilities, all operating on the user's machine.

Voice cloning from audio samples as short as 3 seconds, maintaining tone, timbre, and accent.
Text-to-speech (TTS) generation using seven distinct engines: Qwen3-TTS, HumeAI TADA, Kokoro 82M, Qwen CustomVoice, LuxTTS, Chatterbox Multilingual, and Chatterbox Turbo.
System-wide dictation into any application via global hotkeys, powered by OpenAI Whisper (Base, Small, Medium, Large, Turbo models).
Integration with AI agents (e.g., Claude Code, Cursor) through a local REST API, enabling spoken responses in cloned voices.
Multi-track timeline editor (Stories Editor) for arranging multiple voice tracks, trimming clips, and mixing conversations for narratives and podcasts.
Audio effects pipeline including pitch shift, reverb, delay, and compression for enhanced audio production.
Local or remote GPU inference support across Metal (Apple Silicon), CUDA, ROCm, Intel Arc (XPU), and DirectML architectures.
Refinement of Whisper-generated transcripts using a local Large Language Model (LLM) to clean 'ums' and self-corrections.
Personality profiles for voices, allowing text to be rewritten or composed in character.

use cases

Who Should Use Voicebox?

Voicebox is designed for a diverse range of users who require local, private, and flexible voice generation and manipulation capabilities. Its open-source nature and comprehensive feature set cater to both technical and creative professionals.

Developers and AI Engineers: For integrating voice capabilities into custom applications and AI agents via its local REST API, building voice-enabled applications, and experimenting with local inference.
Podcast Producers and Content Creators: For generating multi-voice narratives, audiobooks, and video narrations using the multi-track timeline editor and high-fidelity voice cloning.
Game Studios: For creating dynamic game dialogue and character voices with precise control over tone and style.
Accessibility Developers and Individuals: For providing speech assistance and accessibility tools, enabling individuals who cannot speak in their original voice to communicate effectively.
Users on Mac with Apple Silicon: For optimal performance and fast voice generation, leveraging the MLX backend for GPU acceleration.

pricing

Voicebox Pricing & Plans

Voicebox operates on a freemium model, with its core application being entirely free and open-source. This model emphasizes local-first operation, eliminating common costs associated with cloud-based AI services. There are no subscription fees, per-token charges, or API rate limits for its local API, providing a cost-effective solution for voice generation and cloning.

Core Application: Free (Open-source, Local-first operation, Voice cloning, Text-to-speech, System-wide dictation, AI agent integration, Built-in REST API, No per-token fees, No API rate limits)

Similar Tools

Voicebox vs Competitors

Voicebox is positioned as a direct, free, and open-source alternative to commercial, cloud-based voice cloning and text-to-speech services. Its primary competitive advantages are its local-first execution, emphasis on privacy, and the absence of recurring costs or usage limits.

ElevenLabsOn Stork

ElevenLabs is a market leader for highly natural-sounding, emotive voice cloning and text-to-speech, particularly for professional audio production.

Unlike Voicebox's local-first and open-source approach, ElevenLabs is a cloud-based proprietary service, offering superior raw output quality for commercial use but with associated costs and data privacy considerations. It operates on a freemium model, but its free plan is limited, and heavy users may find it expensive.

Chatterbox (by Resemble AI)On Stork

Chatterbox is a high-performance, open-source text-to-speech (TTS) model family built for real-time generative audio, offering speed, expressiveness, and zero-shot voice cloning with emotion control.

Similar to Voicebox, Chatterbox is open-source and developer-focused, allowing local deployment and emphasizing real-time performance and expressiveness. It offers a permissive MIT license for commercial use and is designed for production-grade applications.

Coqui TTS (XTTS-v2)On Stork

Coqui TTS, specifically the XTTS-v2 model, is a widely adopted open-source voice generation model known for high-quality, multilingual voice cloning from minimal audio samples.

Like Voicebox, Coqui TTS is open-source and supports local deployment, with a strong focus on voice cloning and multilingual capabilities. However, it is computationally intensive, often requiring a good GPU, and its XTTS-v2 model is available under a non-commercial public model license, unlike Voicebox's MIT license.

MyShell (OpenVoice)↗

MyShell offers OpenVoice, an open-source instant voice cloning AI library that provides unparalleled precision and granular control over tone, emotion, accent, rhythm, and intonation.

MyShell's OpenVoice is an open-source voice cloning solution, similar to Voicebox's offerings, designed for high flexibility and resource efficiency in voice cloning. While MyShell also provides a web app, OpenVoice is primarily an open-source library for developers, emphasizing customization and fine-grained control over generated speech.

See every Voicebox alternative, compared→