Loromatic Support · Iterative Key LLC

Common questions

How do I make apps speak with Loromatic?

Open Loromatic and tap Text-to-speech settings (or go to your phone’s Settings → Accessibility → Text-to-speech output). Choose Loromatic as the preferred engine. Apps that read aloud now use its voices. Tap Listen to an example there to hear it.

Which voice do apps hear?

On the Voices tab, the star marks the default. One of the ten built-in voices is the default for all 31 languages; a star on one of your own voices makes it the default for its language instead. Apps that let you choose a voice can still pick any of them.

Which languages can I clone a voice in?

English, Spanish, French, German, Italian and Portuguese. English works right away; the others need their engine, which Loromatic downloads through Google Play when you ask (165 MB, or 390 MB for a Large engine). After that, everything works offline.

The cloned voice doesn’t sound like the person

Record in a quiet room, about a hand’s width from the phone, reading naturally.
On the review screen, compare the takes: each comes from a different engine. Large engines are slower but usually closer. Try again makes new takes; each sentence comes out a little differently.
Download another engine for the language to get one more take.

Does it work without internet?

Yes. Loromatic has no internet access at all. Only the optional cloning engines are downloaded, once, by Google Play.

Train a voice of your own

Loromatic imports VITS voices trained with Piper, in any language. Because the app doesn’t include espeak-ng, the voice must be trained on text characters instead of phonemes:

Prepare a dataset of your recordings (one or two hours of clean speech gives a good voice; 20 minutes can work when fine-tuning an existing voice of the same language), in Piper’s LJSpeech layout.
Preprocess it with Piper’s training tools using --phoneme-type text.
Train or fine-tune, then export the checkpoint to ONNX. You get voice.onnx and voice.onnx.json.
Put both files in a .zip (optionally with an loromatic.json naming the voice and its language, for example {"name": "Grandpa", "language": "es-MX"}).
Copy the zip to your phone and, in Loromatic, tap Voices → Import a voice.

Voices already prepared for sherpa-onnx (a model with tokens.txt, and a lexicon.txt when the voice uses one) can be imported the same way. Voices that need espeak-ng phonemes without a lexicon can’t be imported; the app explains why when you try.

Terms of use

Loromatic asks you to accept these terms the first time you open it (they are also in Settings → Terms of use, in English, Spanish and French). They include the use restrictions that come with the built-in voices’ license.

These terms are an agreement between you and Iterative Key LLC (“we”, “us”) about your use of the Loromatic app (the “app”). By tapping “I agree”, or by using the app or its voices, you accept them. If you don’t accept them, don’t use the app.

THE APP

We give you a personal, non-exclusive, non-transferable license to install and use the app on devices you own or control, under these terms and the store’s terms. The app turns text into speech on your device. It does not send what it reads, your recordings or your voices anywhere; see our privacy policy at https://iterativekey.com/apps/loro-matic/privacy/.

THE VOICE MODELS AND THEIR LICENSES

The app includes voice models made by others, each under its own license. You can read every license in the app (Settings > Open-source licenses).

a) Built-in voices: Supertonic 3, by Supertone Inc., licensed under the BigScience Open RAIL-M License. The model and its derivatives are subject to paragraph 5 of that license. Its use-based restrictions, reproduced in the appendix below, are part of these terms: you agree not to use the built-in voices, or any output or derivative of them, for any of the restricted uses listed there, and to require the same of anyone you let use them through your device.

b) Voice cloning: Pocket TTS, by Kyutai, licensed under the Creative Commons Attribution 4.0 International license (CC BY 4.0). Nothing in these terms limits the rights that license grants you over those models.

c) Voices you import are yours or their authors’: you are responsible for having the right to use them.

CLONING A VOICE

You may record or import someone’s voice only where the law allows it. You are responsible for having every consent and right the law requires to record, clone and use a person’s voice, including your own recordings of others. The app asks you to confirm this before every recording.

WHAT YOU MAKE WITH THE APP

You are responsible for the speech and audio you generate and how you use them. Don’t use the app, or audio made with it, to break the law, to infringe anyone’s rights, to deceive, defraud or harass anyone, or to pass generated speech off as a real recording of a person. When you share audio made with a built-in voice, say that it is machine-generated (see restriction (e) in the appendix).

NO WARRANTY

The app and its voices are provided “as is”, without warranties of any kind, to the extent the law allows. Generated speech can be wrong, mispronounced or unlike the original voice; don’t rely on it where errors could cause harm.

LIMITATION OF LIABILITY

To the extent the law allows, we are not liable for indirect, incidental, special or consequential damages, or for any loss arising from your use of the app or of audio made with it. Nothing in these terms limits rights you have as a consumer that cannot be limited by contract.

CHANGES AND ENDING

We may update these terms; the app will ask you to accept a new version before you keep using it. You can stop using the app at any time by uninstalling it. If you break these terms, your license to use the app ends.

CONTACT

https://iterativekey.com/apps/loro-matic/support/

APPENDIX: USE RESTRICTIONS OF THE BUILT-IN VOICES (Attachment A of the BigScience Open RAIL-M License, as published with Supertonic 3; “the Model” means the built-in voice model)

You agree not to use the Model or Derivatives of the Model:

(a) In any way that violates any applicable national, federal, state, local or international law or regulation;

(b) For the purpose of exploiting, harming or attempting to exploit or harm minors in any way;

(d) To generate or disseminate personal identifiable information that can be used to harm an individual;

(e) To generate or disseminate information and/or content (e.g. images, code, posts, articles), and place the information and/or content in any context (e.g. bot generating tweets) without expressly and intelligibly disclaiming that the information and/or content is machine generated;

(f) To defame, disparage or otherwise harass others;

(g) To impersonate or attempt to impersonate (e.g. deepfakes) others without their consent;

(h) For fully automated decision making that adversely impacts an individual’s legal rights or otherwise creates or modifies a binding, enforceable obligation;

(i) For any use intended to or which has the effect of discriminating against or harming individuals or groups based on online or offline social behavior or known or predicted personal or personality characteristics;

(j) To exploit any of the vulnerabilities of a specific group of persons based on their age, social, physical or mental characteristics, in order to materially distort the behavior of a person pertaining to that group in a manner that causes or is likely to cause that person or another person physical or psychological harm;

(k) For any use intended to or which has the effect of discriminating against individuals or groups based on legally protected characteristics or categories;

(l) To provide medical advice and medical results interpretation;

(m) To generate or disseminate information for the purpose to be used for administration of justice, law enforcement, immigration or asylum processes, such as predicting an individual will commit fraud/crime commitment (e.g. by text profiling, drawing causal relationships between assertions made in documents, indiscriminate and arbitrarily-targeted use).

Privacy

How Loromatic handles your data is described in its privacy policy.