Pricing

Priced as a delivered outcome. Deployed as software you control.

Avcel is not a subscription to someone else's cloud. It ships in two planes, and both of them answer to you, for regulated institutions and everyday AI adopters alike.

Deployment

Two planes. One perimeter.

Data plane

Avcel Desktop

A native application on each developer's machine. It works with local files and local context directly: nothing is uploaded to a vendor cloud, because there is no vendor cloud.

  • Runs on the developer's own workstation
  • Local files and repositories stay local
  • Connects only to your Control Plane
Control plane: inside your perimeter

Avcel Control Plane

The policy engine, blueprint generation, and audit record run on your own Kubernetes, inside your network boundary. Configure a local model and the deployment has zero external egress. Your models, your keys, your logs.

  • Self-hosted on your infrastructure
  • Default-deny policy, cascading to every project
  • Audit record accumulates where your regulator can see it
Start

Architecture workshop

A half-day with our engineers mapping your constraints to the reference architecture. You keep the blueprint, whatever you decide next.

Deliver

Avcel delivery engagement

Installation, integration, and handover of both planes inside your perimeter, with your first governed workloads delivered in weeks, not a year-long integration project.

License

Software license, your terms

Avcel Desktop and Control Plane are licensed to run in your environment, not rented from ours. Procurement through your standard channels, including an AWS Marketplace private offer if you want to draw down an existing AWS commitment.

Engagement pricing depends on environment, integration surface, and workload scope: we'll give you a real number after the workshop, not a rate card before it. We price delivery, not hours: fixed scope, agreed before work starts, not a time-and-materials clock running against you. Most teams find governed AI delivery costs less than the token bills and integration overruns of doing it themselves.

Common questions

Pricing and procurement questions

How much does Avcel cost?

Engagement pricing depends on your environment, integration surface, and workload scope. AVM prices delivery rather than hours: scope is fixed and agreed before work starts, so there is no time-and-materials clock. You receive a real number after the architecture workshop rather than a rate card before it.

What is the architecture workshop?

A half day. Roughly 90 minutes mapping your constraints (perimeter, policy, examiner posture, data rules), then 90 minutes on how the reference architecture lands in your environment, including gaps and sequencing. You keep the resulting governed AI SDLC architecture document regardless of what you decide next.

Can we buy Avcel through AWS Marketplace?

Yes. AVM is an AWS Advanced Partner and Avcel is available via AWS Marketplace private offer, which lets you draw down an existing AWS consumption commitment and moves procurement through a channel your organization has already approved.

Do we license Avcel or is it delivered?

Both, in sequence. AVM delivers the installation, integration, and handover of the full governed lifecycle inside your perimeter, and Avcel is then licensed to run in your environment rather than rented from ours. You own the running system.