CortexQACortexQA
Pricing, but make it agentic

CortexQA Pricing

Our pricing model is currently training. It wakes up every few epochs and insists it can optimize cost, latency, and flake rates simultaneously.

Translation: we’re not selling it yet. We’re finishing the last mile to make the product worthy of your pipelines.
Context‑Aware Credits

Compute, model, and runner minutes that adapt to your ACP/AFP profiles.

Enterprise Guardrails

SSO/SAML, RBAC/ABAC, Policy‑as‑Code — standard in all paid tiers.

Usage with Budgets

Caps for model cost and latency; route for SLAs with fallbacks.

Founder Access

Early adopters get migration help, custom agents, and priority features.

Join the waitlistTalk to us
Billing
Community
For explorers and personal projects.
$0
/mo
  • Local-first
  • Sample agents
  • Public docs
Join waitlist
Team
For squads shipping with agentic QE.
$110
/mo (billed yearly)
  • ACP/AFP tooling
  • Model router
  • Self-healing
  • Dashboards
Notify me
Enterprise
For regulated, multi-tenant environments.
Custom
  • SSO/SAML/OIDC
  • RBAC/ABAC
  • OPA guardrails
  • Private registry
Talk to us

Play with an estimator

Agent Minutes1200 min
Model Tokens10 M
Runner Capacity20 hrs
Estimated
$64
Not a quote. Agents still negotiating with the CFO.
CFO Smile Index
87%
Higher is better. Keep budgets and outcomes balanced.
Credits
40 AM • 10 MT • 20 RC
AM=Agent Minutes, MT=Model Tokens(M), RC=Runner Capacity(hrs)
We’ll publish real pricing soon. This widget is for vibes, not invoices.

What’s included

Community
Team
Enterprise
Local-first
ACP/AFP
Model Router
Self-healing
Dashboards
SSO/SAML/OIDC
OPA Guardrails

Founder access perks

Founders credits
Free credits for flake-busting & a11y fixes.
Early pricing
Lifetime discount for first adopters.
Flexible billing
Annual discounts, budget caps, usage alerts.
Priority roadmap
Your use-cases shape our next sprints.

A signal from the near future

Our pricing brain is a small swarm of cooperative agents searching the Pareto frontier of cost, latency, and quality. It keeps two promises: make CFOs smile, and never punish good tests.

The emerging shape looks like transparent tiers based on Agent Minutes, Model Tokens, and Runner Capacity — with free credits for flake‑busting and accessibility fixes.

Agent swarm telemetry
loss
0.160
latency_ms
650
flake_ratio
0.030
pareto_points
42
guardrails_pass%
96
budget_drift
0.005

As soon as loss stabilizes and guardrails sign off, we’ll publish the plans. Until then, founder access is open for teams who want to ship with us.

ETA: Soon (like, real soon)