arrow_backBack to the overview
Pricing & packaging

Bought against a headcount line.

The unit of value is a worker, not a seat.

A virtual worker is evaluated against what a competent person in that role produces, which is why the objective ladder, the outcome contract and the improvement loop matter more than model quality. It also means the number that matters in a business case is projected cost per worker per month, and the platform is built to tell you that number rather than leave you to estimate it.

One motionthe same structure as the Knowledge Substrate
Auditablea usage meter you can reconcile
Capped pilotsa first class configuration, not an exception
The layers

Seven things you can buy.

Pricing is quoted per engagement, because worker count, channel mix, language coverage and conversation volume vary too widely for a list price to be honest.

LayerWhat you getBasis
Hub platformThe runtime, the channel fabric, the calibration subsystem, the control plane, the studio, and the governance and compliance regimePlatform licence, per installation
Worker packsA commissioned worker: definition, agent template, ladders, flows, guardrails, outcome contract, scoring rubric and evaluation suitePer worker, per seat equivalent, or per outcome
Templates, voices and skillsCurated personality, per language voice mappings and licensed skill packs beyond the standard libraryIncluded at base, premium and bespoke chargeable
Domain contentThe licensed knowledge that makes a worker competent in a named domainSubscription, on the Knowledge Substrate model
Channel, model and usageTelephony, messaging and model consumption, priced per task classPass through plus margin, against an auditable meter
ServicesCommissioning, knowledge and asset onboarding, methodology capture, calibration tuningFixed scope engagements
Billing and taxHandled by Cleverbridge as merchant of record: payment, tax calculation and remittance, invoicing, dunning and refundsNo per jurisdiction tax registration on your side
Engagement shapes

Three ways to start.

Each is an entitlement configuration rather than a manual arrangement, which is what allows a pilot to be provisioned quickly and converted without a rebuild.

Start here

Capped pilot

A bounded proof against your own accounts, sized so the result is meaningful and the exposure is not.

  • checkOne worker, one channel, one language
  • checkA fixed simulation allowance and a spend ceiling
  • checkTime bound, with a defined success measure agreed up front
  • checkFull telemetry and cost reporting from the first conversation
  • checkConverts without a rebuild
eventScope a pilot
Most common

Worker deployment

A commissioned worker running against your live pipeline, supervised by your team, improving on evidence.

  • checkPlatform licence plus one or more worker packs
  • checkVoice, messaging, email and web chat
  • checkBidirectional integration with your system of record
  • checkCalibration, simulation and the request board
  • checkRole based dashboards and the control plane
  • checkDomain content subscription where a licensed domain applies
eventSchedule a briefing
At scale

Fleet

Several worker types across teams and markets, with supervision, entitlement management and internal facing workers.

  • checkMultiple worker packs and templates
  • checkMulti-jurisdiction compliance configuration
  • checkFleet supervision, human or virtual
  • checkEntitlement, budget and cap management per team
  • checkMarketing and support workers serving the board
  • checkDeployment in your environment, or Corvair operated
forumTalk to us
Packaging & distribution

Metered in credits and minutes, never in our cost.

Two meters, both yours to see. Voice minutes, because that is the unit you already budget in and the one carriers bill. Credits, a single currency covering every other metered activity.

What each activity costs in credits is set by the publisher of the worker pack and ships inside the pack itself. What a credit costs in money is set by whoever sells you the deployment. That separation is deliberate: it is what lets a white label, a marketplace listing and a Corvair-operated tenant price differently without anyone forking the engine.

Because the schedule ships in the pack and versions with it, your rate cannot move underneath you. A repricing arrives as an upgrade you approve, itemised against your current schedule, rather than as a change of terms you discover in an invoice.

Publisher credit schedule listing activities and their credit cost, alongside pack contents.
The credit schedule, as the publisher sets it. A connected voice conversation and each voice minute, messaging per thread per active day, research passes, match runs, asset generation, enrichment lookups and simulation cells. The schedule is a file in the pack, versioned with everything else in it.

The schedule is a design instrument, not just a price list. In the property pack an enquiry callback is priced below a general conversation, deliberately, because speed of first callback is the metric that moves performance most and the schedule should not discourage the ten-minute callback. An appraisal lead call is priced the same way, because winning the instruction is the most valuable outcome the agency produces.

The three ceilings behave differently, and the differences matter more than the numbers. Minutes exhausted: conversations in progress finish, no new outbound dial, and text channels continue if credits remain. In property, inbound enquiry callbacks are the last thing to stop, because an unanswered enquiry is a lost viewing. Credits exhausted: discretionary consumption stops first, being simulation, research refreshes, enrichment and match runs, while scheduled work is honoured and the queue holds rather than drops. Licence ceiling: a further worker cannot be deployed, existing workers are unaffected, and the ceiling is visible before provisioning rather than after.

What you never see on your invoice is our cost of goods. Model tokens and audio seconds are our problem, and a platform that shows you its margin instead of your bill has confused two different products.

Entitlement suspends; it does not degrade. A worker at zero credits stops and says so. A worker that quietly gets worse as the balance falls is the least debuggable failure we could ship you.

The unit is a worker pack.

A worker definition is data, not code. A worker pack is that data made portable, versioned and signed, and it is what ships in every model below.

What is in a pack

The worker definition, the template and application prompt layers, the agent template and voice mappings, ladders, flows, guardrails, the outcome contract schema, the scoring rubric, the skill and tool manifests, the knowledge scope declaration, the connector requirements and the entitlement shape.

What is never in one

The kernel prompt layer, which is engine code and is what makes every worker in the family behave like one. And customer data, in any form, ever.

science
The evaluation suite travels with the pack. A worker you can acquire that arrives with the tests proving it works, runnable in your own environment before you deploy it, is a different object from a prompt in a zip file. It turns our promotion gate into an acceptance test you run rather than an assurance we give, and it is the only honest way to sell a worker to somebody who has not watched it work.
key
The declared authority profile makes a pack legible before it runs. A pack cannot grant itself anything. It declares which privileged activities it needs and at what grade, the way an application declares permissions, so you see that a worker wants discounting authority within a band, or access to compensation data, and you grant, narrow or refuse. A pack requesting nothing privileged can be tried without a conversation.
Deployment modeWhat shipsWhose brandWho operates itWho is accountable for compliance
DirectA tenancy on the Corvair platformCorvairYouYou
Corvair operatedA configured worker we run on your behalfEitherCorvairYou, with Corvair as processor, stated in the contract
White labelThe engine in your own cloud project, under your brandYoursYouThe distributor
EmbeddedA single worker pack running inside your productYour productYouThe distributor, unless the host passed it on in writing
MarketplacePacks, bundled with domain content, data or API entitlementsPublisher branded in a catalogueWhoever installsThe tenant
tune
The mode is not only a commercial shape. It fixes who may edit what. Setting it is the first thing an onboarding does, because it determines the default edit authority on every prompt layer for every subject, and running it later means re-deciding choices already made under the wrong assumption. Under white label and embedded, a distributor holds none by default on the template and application layers, and Corvair may grant overlay per partner, per layer, with a stated scope and an expiry. Where a partner hosts, Corvair locks the prompts. Grants are revocable, and a revocation does not unwind live overlays: the partner loses the ability to author new ones, and rolling back a specific overlay is a separate, deliberate decision.

Not a roadmap in order. Five shapes the same engine has to fit. White label turns brand into configuration and needs a stated version support boundary, because a partner runs a version rather than the current one. Embedded workers frequently use the host product's own interface as their channel, which costs nothing architecturally because a channel is an adapter. A marketplace sells certification, not packs: schema and guardrail conformance, the pack's own evaluation suite executed with published results, a compliance review of its declared purpose and data classes, and an authority profile checked for proportionality. A lead qualification worker requesting compensation data does not pass.

shield
Four kernel floors, and what happens to each. A floor rule is one whose absence makes a compliance claim false: the AI disclosure, the identity resolution rules, the suppression handling and the no-invented-facts rule. Rewording one, for instance to a jurisdiction's mandated disclosure phrasing, is an overlay: it takes engineering approval, full re-certification for every jurisdiction, and a note on the deployment's compliance record. Adding to one or narrowing what it permits is the same. Removing one is not an overlay, it is a variant, with its own pack lineage, its own signature and its own certification, and the deployment is no longer running the standard kernel. Removal is available. Recording it as an overlay would make the audit trail say something untrue.
lock
What no deployment may disable. The kernel prompt layer stays closed. The compliance engine stays in the request path and is not a middleware a partner may remove. Authority enforcement holds, so a pack requests grades and never grants itself one. Suppression propagates immediately in every model. The audit trail is append only. The meter exists everywhere, because an unmetered deployment is an unbillable and unaccountable one. And one that matters more than it sounds: a worker answers truthfully when asked whether it is automated, under every brand. Brand is configurable. That is not.
extension
A pack may extend, never weaken. It adds ladders, flows, knowledge scope and skills. It cannot lower a guardrail, widen a data class, disable a check or raise its own authority, and the engine validates that on install rather than trusting the manifest, because a manifest is a claim and installation is where claims get tested.
The argument

Two reasons this is defensible.

The category

The market has lead generation tools and it has call automation tools. It does not have a governed runtime for role holding workers. A buyer does not ask which model you use. They ask what it achieved, how you know, and what happens when it gets something wrong. The Hub has an answer to all three, and each answer is an artefact rather than an assurance.

The compounding asset

Every conversation your workers hold makes them measurably better, and the improvement is expressed in versioned artefacts you own and can inspect: prompts, knowledge, assets and skills. That is a switching cost built from delivered value rather than from lock-in, and it is why the calibration loop is priced as part of the platform rather than as an add-on.

savings
On usage costs. Audio dominates conversational spend by a wide margin, which is why the conversation policy that makes a worker sound like a competent colleague, being one short sentence per turn rather than a monologue, is also the largest cost lever in the system. Quality and economics point the same way here, and the platform reports both.
Next step

Start with a conversation, not a contract.

The most useful first session is a working demonstration against a scenario from your own pipeline, followed by an honest scoping of what a pilot would involve.