Pricing

€ 0,07 a minute, and you can see everything under it

One orchestration fee. Model usage passed through at cost on our keys, or billed to you directly on yours. No per-seat licence, no markup hidden inside the model calls.

applies to inbound, outbound and in-app web calls alike

What you pay

Four ways to run the model layer

The orchestration fee is the same in every mode. What changes is who buys the models and who gets billed for them.

ModeWhat you payWhen it fits
Platform keys (default)Model usage at cost + € 0,07/minYou want to start today. Every supported provider works without you opening an account anywhere.
Bring your own keysYour provider bills you + € 0,07/minYou want to negotiate volume yourself, or keep provider-level control over how data is handled.
Gemini realtime via Vertex AI€ 0,25/min all inYou want one predictable number per minute and native-audio conversation quality.
Self-hosted models, by arrangementFree + € 0,07/minOpen models running on your own infrastructure. Not switchable from the dashboard: it needs a conversation, a licence and a deployment we set up with you.
The first three modes are available the moment you create an account. The fourth is custom work: see what is self-serve and what we arrange together, below. Telephony is billed separately, per leg, at the carrier's rate. A bridged transfer holds two legs for its duration.

Self-serve or together

What you can start alone, and what we do with you

We would rather be clear about this before you sign up than after. Creating an account gets you the whole platform for calls that run on our infrastructure. The heavier things (models on your own hardware, an on-premise deployment, the compliance file) are not features we are withholding behind a plan. They are the parts that only work when they are built for your environment, and that is the work we do alongside your team.

The moment you create an account
Every supported STT, LLM and TTS provider on our platform keys, at cost
Inbound, outbound and in-app web calls on the same pipeline
The dashboard and the full REST API, with your own API key
Live transcription, real-time call control and webhooks
Post-call analysis against a schema you define per assistant
Bring your own provider keys whenever you want to switch
Prepaid credit, so you can start without a contract
By arrangement, with us
Local and self-hosted models, scoped per use case, under a licence
On-premise, private-cloud or air-gapped deployment, with a licence
Trust Center access: the compliance and security documentation, DPA and subprocessor list
White-label and reseller setup, including per-organisation pricing
Invoicing instead of prepaid credit
A custom SLA, and security review with your own team

Why the second column is not a paywall

A voice agent on your own hardware, or inside an air-gapped network, is not a checkbox. It is a deployment with your telephony, your models and your security requirements in it. If we let people switch that on unattended, most of them would end up with something that technically runs and does not work. Doing it together means we can actually give it attention: we design it around your environment, debug the telephony at the log level, and stay on the layer underneath afterwards. That is not an upsell, it is the whole reason VoiceDock exists.

In the fee

What the € 0,07 actually covers

The orchestration fee is not a licence to access an API. It is the layer that holds the call together.

Recording and storage are configurable per organisation, including a mode that stores nothing.

Session handling, audio routing and turn-taking for the whole call
Inbound, outbound and browser-based web calls on the same pipeline
Live transcription and real-time call control while the call is running
Webhooks for tool calls, status changes and the end-of-call report
Post-call analysis against a structured schema you define per assistant
The dashboard and the full REST API, for you and your team

The build

The platform is the cheap part

Most of the value in a voice system is in the design of the conversation, the tool calls into your own systems, and the telephony behaving under load. That work is scoped per engagement rather than priced per minute, and we do it with your team rather than for them.

Working sessions with the engineers who own the orchestration layer
Call flow, tool calls and provider stack designed around your use case
Live deployment, with telephony debugged at the log level
Full handover: dashboard, API and documentation, no lock-in

Scoped per engagement. Built for teams ready to put a real phone line on this, not a weekend prototype.

Talk about a build

Enterprise and sovereign

When the constraint is regulatory

Private cloud in the EU region of your choice, or a full on-premise installation on your own hardware including air-gapped networks. This runs under a licence and is priced per deployment, because the work depends entirely on your environment. It starts with a conversation, not a sign-up.

Questions

The things worth asking any voice vendor

What do I actually get the moment I create an account?

The full platform on our infrastructure: every supported provider on our keys at cost, inbound, outbound and web calls, the dashboard, the REST API, webhooks and post-call analysis, on prepaid credit. What you do not get automatically is anything that lives outside our infrastructure or outside a contract: local and self-hosted models, an on-premise or private-cloud deployment, Trust Center access to the compliance documentation, invoicing, and white-label setup. Those are arranged with us, under a licence where one applies.

Why can't I just switch on self-hosting myself?

Because it is a deployment, not a setting. It involves your hardware, your telephony, your models and your security requirements, and switching it on unattended tends to produce something that technically runs but does not work on a real phone line. We would rather build it with you, which is also how we can keep supporting it afterwards.

Is model usage marked up?

No. On our platform keys, provider usage is passed through at cost, which means the € 0,07 orchestration fee is the whole of our margin on a minute. That is the single question we would ask any voice platform first, and it is why we answer it in public.

Does the meter run during silence and hold?

A caller who is still on the line is still an open session, so yes, orchestration is billed on the duration of the call rather than on speech alone. It is worth knowing because it also means shorter, better-designed calls cost less.

How is a transfer billed?

A bridged transfer holds two telephony legs for as long as the conversation continues, so both legs are billed by the carrier. A blind transfer hands the call away and ends our leg, which is cheaper and gives you no visibility into what happened next.

Do I need a minimum volume?

Not for the platform. The co-engineering work is scoped per engagement, so there is a practical floor on that side. If your volume is a handful of calls a month, a partnership is not what you need and we will say so.

Can I pay on invoice instead of prepaid credit?

Yes. Prepaid credit is the default and suits most teams; invoicing is available per organisation, which is the normal arrangement for custom builds and reseller setups.

What happens if my own provider key stops working mid-call?

We fail closed. The call ends or is refused with a reason recorded against it, rather than being silently switched to a provider you did not choose. A call that quietly behaved differently than designed is harder to trust than a call that stopped honestly.

Can I resell this under my own brand?

Yes. A parent organisation can cover usage for its child organisations, with per-organisation pricing multipliers, so you set your own price to your own customers. This is how the white-label and agency setups run.

Bring us a number you want answered properly

Thirty minutes. We map the call flow, name the provider stack, and give you a cost per connected conversation you can put in a spreadsheet.