Docs Book a Demo Sign in
Enterprise

More calls, your servers, or your device.

When the self-serve product isn’t enough: more calls at once, a deployment of your own, on-premise, or our models inside your product. The people who build the models answer.

What we do

Tell us which one.

  • More calls at once Accounts take 10 calls at once after the first top-up. Tell us your peak and we raise your limit where shared capacity allows, or give you capacity of your own. Ask about this →
  • A dedicated deployment Your own instance of the stack, with capacity no other customer shares, in our cloud or in yours. Ask about this →
  • On-premise The full pipeline on your own servers. Audio never leaves your infrastructure. For healthcare, finance and the public sector. Ask about this →
  • On your device The whole voice agent inside your product, offline: our full models, licensed and ported to your hardware. Ask about this → On-device
  • Voices and languages A voice of your own from a recording, or a language we don’t speak yet. Ask about this →
  • Volume and paperwork Pricing for committed volume, invoices for your finance team and answers for your security review. Ask about this →
Contact

Send an enquiry.

We reply by email. Prefer to talk? Book a call or write to contact@lokutor.com.

What do you need?

FAQ

Common questions.

How many calls at once can a Lokutor account take?

10 after its first top-up and 3 before. For more, tell us your expected peak through the form on this page: we raise an account’s limit where our shared capacity allows it, or set up a dedicated deployment.

Can Lokutor run on-premise?

Yes. Our models are small enough to run in real time on an ordinary CPU, so the whole voice pipeline can run on your own servers and audio never leaves your infrastructure. On-premise deployments are scoped and quoted with you.

Do you license your models for devices?

Yes. Our full models can run inside a product with an application processor, such as a robot, a vehicle or an appliance, so the whole voice agent works offline: commercial licences, porting to your hardware and models for other languages. For microcontrollers there are Ito and Oído, built for a $5 chip.

Where does Lokutor process data?

Our cloud runs in the EU: calls are processed in AWS Ireland and accounts and conversation records are stored in Frankfurt. The language model comes from a provider. On-premise and on-device deployments keep all audio inside your own infrastructure or device.