Is the voice forge really free?

Yes for voices designed on your machine with Qwen3-TTS VoiceDesign. You download the model once. There is no meter and no subscription for that path. It runs on Windows x64 today.

Can I make a face from a photo for free?

Bundled humans, animals and robots are free. A face built from your own photo needs an Engramm license. Creating several avatars from the free faces stays free.

Does it clone a real person?

The forge designs a voice from a written description. Cloning from a short sample you record is a separate local path on Qwen3 (Windows). Do not feed another service's voice outputs into it.

Settings › Profile has the button, and the dock can open Avatar Studio. Inside, Create a voice opens the forge on the Voice node.

Avatar Studio · 2.0

A face, a voice, a personality. One avatar per desktop.

Open Avatar Studio from the board. Pick a face from the catalogue, design a voice on your machine for free, and wear a different assistant on every desktop.

Download Mnemosyne OS All features

Avatar Studio

Your assistant becomes someone you can see.

An avatar is a name, a face of particles, a voice and a language. From Settings › Profile or the dock, open Avatar Studio: a catalogue of humans, animals and robots, then an atelier with the avatar at the centre.

The face breathes, blinks and follows sixteen emotions in sync with speech. Built-in faces stay free. A face made from your own photo needs an Engramm license.

Avatar Studio in the manual

Voice forge

Describe a voice. Keep the take you like. Local and free.

The voice forge sits inside Avatar Studio. You describe the voice (person, animal or robot), hear three takes, and keep the best one under a name. The catalogue of voices you made stays on your machine.

The engine is Qwen3-TTS VoiceDesign (Apache-2.0). It invents a voice from the description. There is no per-voice fee and no cloud bill for this path. On Windows x64 for now; other platforms keep Piper, system voices and ElevenLabs.

Desktops

Work wears one assistant. Home wears another.

Each desktop of the board can wear its own avatar. The backdrop follows the desktop you show. Settings › Voice keeps the app-wide voice; the avatar voice speaks while its desktop is active.

FAQ

Questions

Is the voice forge really free?: Yes for voices designed on your machine with Qwen3-TTS VoiceDesign. You download the model once. There is no meter and no subscription for that path. It runs on Windows x64 today.
Can I make a face from a photo for free?: Bundled humans, animals and robots are free. A face built from your own photo needs an Engramm license. Creating several avatars from the free faces stays free.
Does it clone a real person?: The forge designs a voice from a written description. Cloning from a short sample you record is a separate local path on Qwen3 (Windows). Do not feed another service's voice outputs into it.
Where do I open it?: Settings › Profile has the button, and the dock can open Avatar Studio. Inside, Create a voice opens the forge on the Voice node.