Our voices are the backbone of our product and we’re proud to offer a variety of voice creation options to suit your needs.

​
Types of voices

LMNT offers two types of voice creation: instant and professional.

​
Instant voices

Instant voices are available immediately in all of your applications. While they may not always be perfect, they are often still of a very high quality. Instant voices are a great way to quickly test out a voice.

​
Professional voices

Professional voices are fine-tuned to sound as natural, accurate, and lifelike as possible. They require some time to fully train. After completion, you will receive an email notification and can start using this voice in all of your applications.

Users are limited to a number of professional voice creations based on their plan.

​
How to create a voice

You can create a voice in our web app or via our API and SDKs.

​
Web app

To create a voice in our web app, simply navigate to the cloning page. Upload a speech sample, give the voice a name, and click “submit”.

Note that our web app only allows professional voice creations at the moment. If you would like to create an instant voice, use the API.

Some tips:

  • Source material should:
    • Contain only one person’s voice, in English
    • Not contain background noise/music
  • Consider Adobe’s speech enhancer if your speech is not professionally recorded
  • If cutting larger source material down:
    • Leave gaps between sentences (good)
    • Or upload sections as separate files (best)

​
API

To create a voice via our REST API, you can use the create voice endpoint.

You can use a curl command like this to create a voice in your terminal:

curl --request POST \
  --url https://api.lmnt.com/v1/ai/voice \
  --header 'Content-Type: multipart/form-data' \
  --header 'X-API-Key: 5125c321ee36420ea9bef2341974e0f2b' \
  --form files=file:///Users/user/Downloads/filename.wav \
  --form 'metadata={"name": "My Voice", "type": "instant", "enhance": true }'
Don’t forget to get your own API Key!

If you choose to use the API, you can create both instant and professional voices.

​
SDKs

We also offer SDKs for Python and Node.js that make it easy to create voices. Learn more about our SDKs here.

​
How to use a voice

Once the voice is ready to use, you can use it to synthesize text.

In the Speech Playground, the voice will appear in your list of voices.

In the APIs and SDKs, you will need the voice id. You can find this in the voice list or in the endpoint.

​
How to delete a voice

Currently, you can only delete a voice using our API or SDKs. You can use the delete voice endpoint to delete a voice by id or use the equivalent method in your SDK.

Optimizing for Low Latency