2¢ a minute. Everything included.
Speech recognition, the language model and the voice, in one price per minute — down to 1.4¢. Elsewhere the platform fee alone is 5–8¢, before you choose a model.
No card. Try it with your own prompts and voices.
Small apps and prototypes.
Adds voice cloning and phone numbers.
High-volume production.
Every plan also includes text-to-speech and speech-to-text API allowances: 100K / 2M / 12M / 30M characters and 60 / 1,000 / 6,000 / 25,000 minutes. Phone numbers $5 / month each. Dedicated instances, on-prem and on-device licensing are quoted separately.
The whole minute, for less than their platform fee.
| Platform | Published price per minute | What that covers |
|---|---|---|
| Lokutor | 1.4–2¢ | Everything: speech recognition, language model, voice |
| Cartesia | 6¢ | Call duration |
| Deepgram Voice Agent | 7.5¢ | Standard tier, pay as you go |
| ElevenLabs Agents | 8¢ | Per minute beyond a plan’s allocation |
| Retell AI | 5.5¢ + LLM + voice | Voice engine; language model and TTS billed on top |
| Vapi | 5¢ + providers | Platform fee; recognition, LLM and TTS billed on top |
Each company’s own pricing page, read 23 September 2026 — follow the links to check. Web calls; telephony is extra on every platform, ours included.
How you’re billed. All of it.
- The minuteYour plan is a monthly fee with minutes included. Speech recognition, the language model and the voice are all inside the minute — there is no other per-minute charge and no model to pick and pay for separately.
- MeasuredPer second of connected session. Unused minutes don’t roll over.
- Phone callsPhone lines carry a carrier cost. A minute of a call your agent receives counts as two (3.3¢ on Growth, 2.8¢ on Business); a minute of a call it makes counts as four (6.6¢ / 5.6¢). Numbers are $5 a month each.
- Web searchOff unless you turn it on for an agent. Each search counts as four minutes.
- At your limitCalls stop until next month — no surprise bill. Or turn on overage and keep going at your plan’s overage rate, up to one extra month’s minutes, billed with your next invoice.
- Changing planUpgrade any time: a new month starts that day, less credit for what’s left of the old one. A smaller plan applies straight away, without a refund.
What would it cost you?
Drag the sliders. We pick the cheapest plan that fits and show the maths.
- Starter plan, 1,000 min included$19
- No extra minutes$0
- No phone numbers$0
Estimates from the public price list. Every minute includes speech recognition, the language model and the voice. A minute of a call your agent receives counts as two, of a call it makes as four; phone numbers are Growth and up, $5 a month each. Overage runs to one extra month's minutes.
Common questions.
What is the cheapest voice AI agent platform?
Lokutor charges 1.4–2¢ a minute with speech recognition, the language model and the voice included, against published rates of 6–8¢ a minute elsewhere, several of which bill the language model and voice on top. The Free plan gives you 60 minutes a month with no card; paid plans start at $19 a month for 1,000 minutes.
Why is it so much cheaper?
Our voice model is a few tens of millions of parameters, not billions, and runs on commodity hardware — a fraction of what the large GPU models behind most voice platforms cost to serve. That is what lets us price the whole minute at 2¢ and run a sustainable business on it. It is not a promotional price.
Which languages are supported?
Spanish, Catalan, Galician, Basque, Portuguese, English, French, Italian and German — including the Iberian languages most voice platforms don’t offer at all.
Can voice AI run on-premise?
Yes. Lokutor's full pipeline — turn-taking, noise suppression, speech-to-text and speech synthesis — runs inside your own infrastructure with audio that never has to leave your environment. See Solutions for on-premise deployment.
What is on-device or edge voice AI?
Running the voice pipeline directly on the compute already inside a product — a device, robot or appliance — instead of round-tripping to a cloud GPU. Because Lokutor's models run on ordinary CPUs, they fit on-device without adding a per-unit cloud inference bill.
What's included in every plan?
Speech synthesis in nine languages, speech-to-text, the language model, orchestration, tool calling and cross-call memory, plus text-to-speech and speech-to-text API allowances. Growth and above add voice cloning and phone numbers.
Start on the free plan.
60 minutes a month, no card. Then 2¢ a minute, all in.