Pricing

One flat price.
Everything included.

No seats to count and no feature held back for a higher tier. One price, per project, per month, and your own AI key if you want it.

If your data, hosting or model has to stay inside your own boundary, there is a custom option, including your own cloud and data residency.

How it starts · with us

Qarmiq is not self-serve yet. We set each team up directly: your environment URLs, a few promises written as plain sentences, and a first run.

Custom

Your infrastructure. Your rules.

For teams whose data, hosting or model has to stay inside their own boundary. Scoped with you and priced to what you need. These are arrangements, not settings you switch on.

Custom scoped with you
Talk to us
Everything in Project +
  • On-premise or private cloud Qarmiq Edge appliance Scoped with you i
  • Your own database Scoped with you i
  • Data residency, set once Scoped with you i
  • Runners in your own AWS account Scoped with you i
  • Self-hosted model i
  • Written terms: data processing and security review Agreed per partner
  • Single sign-on Roadmap

There is no trial to start, because we set each team up directly. The demo is the way in. Anything marked “scoped with you” is agreed per customer, and anything marked roadmap is not available today.

Deployment

Where it can run.

The flat price covers hosted. The other two options are for teams whose model or data cannot sit on someone else’s infrastructure.

01 · Hosted

We run it

Qarmiq hosts the platform and we set your project up with you. You add environment URLs and, for pages behind a login, test-account sign-ins, then write promises and start a run.

The flat $499 per project, per month.
02 · Your AI

You choose the model

Give a project a key from OpenAI or Anthropic, or point it at a model you host. Your usage runs on your key, and you can see it by feature and by month.

Included at the flat price. How it works
03 · Your environment

It runs where you do

Your own database, your own runners, a private cloud or on-premise deployment, and a data residency you chose once. Scoped with you, because each one depends on where your data has to stay.

Custom. Where data lives

Most teams need only the first. If your requirements point to the third, tell us what has to stay where and we will tell you straight what is built, what is scoped per customer and what is roadmap.

BYOAI

Bring your own AI.

Every project can run Qarmiq’s AI features on a model you control, so the spend and the prompts sit with you.

Your provider

OpenAI, Anthropic, or your own

Set per project in its integrations settings: a key for OpenAI or Anthropic, or an OpenAI-compatible model you host yourself. The key is stored encrypted.

Your cost

Your key is not capped

A call uses the project’s own key when one is set, and falls back to the platform default only within the same provider. Platform-default calls are capped each month. Your own key is not.

Your visibility

Tokens by feature, by month

An AI usage page shows what each feature used, so the cost of a verdict, a brief or a stabilization plan is something you can read rather than estimate.

Data residency

Where your data lives is a decision, not a surprise.

On Qarmiq’s own hosting we tell you where your data sits before anything starts. If it has to live somewhere specific, that is the custom option, and these are its rules.

Decided once

Residency is chosen when your organization’s first project is created, then locked. There is no migration tool, on purpose, so data cannot drift from where you said it lives.

Never silently moved

If your own database is unreachable, you get a hard, visible error. Qarmiq never falls back to the shared database to keep things working.

Only the testing data

Scenarios, checks, runs, verdicts, defects and insight aggregates move to your database. Sign-in, organization and membership records stay in Qarmiq’s core database, so login keeps working.

Bring-your-own database and infrastructure are designed and agreed per customer. They are not generally available, and not switches in a settings page.

Common questions

What people ask first.

Is it priced per seat?
No. The price is per project, per month. It does not change when your team grows or a new person joins, and no capability is held back for a higher tier.
What counts as a project?
One product under test, with its own scenarios, environments, runs and members. If you test two products, that is two projects.
How does it start, and is there a trial?
There is no self-serve trial, because Qarmiq is not self-serve yet. We create the project with you, add your environment URLs and any test-account sign-ins, write your first promises and start a run. Witness needs one script tag in your app, and we set up the test runner for autonomous testing.
Does the AI cost extra?
Platform-default AI calls are capped each month. If you bring your own key, your usage runs on your provider account instead and is not capped by Qarmiq, with an AI usage page that shows tokens by feature and month.
Can it run on-premise or air-gapped?
That is the custom option, scoped with you. Running in your own cloud, using your own database and runners, and a self-hosted model for the AI side are designed and agreed per customer. They are not a setting you can switch on today.
What about terms, data processing and compliance?
Qarmiq has not published a privacy policy, terms or data processing agreement yet, so these are agreed in writing with each partner before you start. We are building toward SOC 2 Type II and are not certified.

Something missing? Email us and a person replies. If Qarmiq is the wrong fit, we would rather say so.

One price. Add the boundary you need.

Start on the flat price and keep it flat as the team grows. If your data or hosting has to stay inside your own walls, we scope that with you, and nothing you have written is lost.

Coding agents connect over MCP as part of the price. The levels page shows where that fits.