Open source · AGPL-3.0 · v1.0

Fuck-off ElevenLabs, and Pay Services.

The voice studio that runs on your own computer. Clone, design, dub, dictate, transcribe and narrate in 646 languages. No account. No subscription. No credits.

Download for Windows All download options

GitHub stars: —
Speech languages: 646
Runs: Offline
Price: $0

ElevenVoice

Clone a voice from seconds of speech.

Record or drop five to fifteen seconds of clean audio. ElevenVoice turns it into a voice you can use in any of 646 languages.
Or start from hundreds of voices.

Pick a ready-made voice from the gallery, or describe a new one in words and design it.
Dub whole folders at once.

Queue videos, choose their languages, save the setup as a rule and let a watched folder do the rest.
Chain it into workflows.

Transcribe, translate, speak and export in one visual flow, or drive everything from the local API.

Clone. Design. Dub.
All of it, free.

MVMy voice12 s sample

Every voice has a story. Bring yours to life.

Synthesize

Clone your voice

Turn a short sample into speech in any language.

A warm, unhurried narrator

Young adultMiddle-agedElderly

Design voice

Design a voice

Describe the voice you want. No recording needed.

Welcome back to the studio.

Bem-vindo de volta ao estúdio.

Bienvenido de nuevo al estudio.

ENPTES

Dub a video

New language, original timing, background kept.

Move the launch to Thursday and send the draft to Sam.

Dictate anywhere

Hold a shortcut, speak, and your words appear as text.

Narrator The storm arrived before dark.

Keeper Light the lamp. Quickly now.

Sailor There, on the water!

Tell a story

Cast a script and give every role its own voice.

0:00 Thanks for making the time today.

0:04 Of course. Where should we begin?

SRTVTTTXT

Transcribe audio

Timed text from any recording, powered by Whisper.

The Lighthouse

1 · The storm arrives

2 · A light on the water

3 · What the keeper knew

Narrate a book

Every chapter read aloud, ready to export.

Browse voices

Hundreds of ready-made voices to start from.

agent Read the build summary aloud

mcp ElevenVoice · speak

Local API on this machine

Connect your agents

REST API and MCP server for Claude, Cursor, n8n and more.

The "Pro" tools, free.

Other studios lock production tools behind a plan. In ElevenVoice they are just features.

Version review: Approve the take you sign off on, keep revision notes and compare two takes side by side.
Batch rules & reports: Save languages, voice and background as a rule, watch a folder, export a CSV of every job.
Pre-delivery check: Untranslated lines, missing tracks, silent or clipping audio and tracks longer than the source, flagged before the client sees them.
Client delivery package: One zip per project: dubbed audio, SRT and VTT subtitles and the timed script for every language.

Appearance settings with the Studio theme

ElevenVoice and ElevenLabs,
row by row.

What is compared	ElevenVoice	ElevenLabs
Runs on	Your computer	Their cloud
Cost	$0, forever	Free tier, then paid credit plans
Account	None	Required
Your audio	Stays local	Uploaded
Usage limits	Only your hardware	Characters per month
Works offline	Yes	No
Source code	Open, AGPL-3.0	Closed
Commercial use	App: yes · voices: each model's licence	Paid plans

Based on ElevenLabs' public pricing page. ElevenVoice is not affiliated with ElevenLabs.

Download ElevenVoice.

The Windows installer sets everything up on first launch. Prefer the terminal? One command installs and opens the app on any system.

Windows 10 / 11 Download the installer v1.0.0 · .exe · 64-bit macOS · Linux All releases Builds and changelog on GitHub

PS> irm https://raw.githubusercontent.com/Hylanderry/ElevenVoice/main/install.ps1 | iex

$ curl -fsSL https://raw.githubusercontent.com/Hylanderry/ElevenVoice/main/install.sh | sh

Prefer to do it by hand? Manual steps on GitHub.

Questions.

Is it really free?

Yes. There are no plans, credits or sign-in. The code is open source under the AGPL-3.0.

What computer do I need?

Windows 10/11, macOS on Apple Silicon or a 64-bit Linux, with 16 GB of RAM recommended. The default voice engine wants a GPU with about 6 GB of memory; lighter engines run on any computer.

Can I use the voices commercially?

The app is free for any use, but each voice model has its own licence. The default engine (OmniVoice) is for personal and research use. Check every model's notice in Settings → Models before paid work, and only clone voices you have permission to use.

Does anything leave my computer?

No. Models download once from Hugging Face; after that everything runs offline. Online translators are optional and only used if you add your own keys.

Windows says the installer is from an unknown publisher.

The installer is not code-signed yet. Click "More info" and then "Run anyway", or install from the terminal command above, which builds from the public source.

If it saves you a subscription, give it a star.

Star ElevenVoice on GitHub