Pricing

Pay for the ceiling, not the chaos.

Metered on seats and routed requests. You keep paying the model hosts directly — we just stop the bill from exploding.

Starter

Solo builders and early products.

$24/mo

  • 2 seats
  • 3 endpoint types
  • 500k routed requests
  • Hard caps + email alerts
  • 7-day trace retention
Start a free week

Scale

Multi-brand, multi-region control.

$319/mo

  • Unlimited seats
  • Identity federation and audit log
  • Custom rate fabrics
  • EU / US pinning
  • Named limits engineer
  • 1-year retention
Talk to Scale
CapabilityStarterTeamScale
Hard monthly caps1 workspaceUnlimitedUnlimited + orgs
Model routingManualPolicy enginePolicy + shadow traffic
Virtual keys5UnlimitedUnlimited + just-in-time
Prompt cache—IncludedIncluded + regional
SupportDocsShared channelNamed engineer

Questions finance will ask

Do you take a cut of our model invoices?

No. You keep billing relationships with the hosts. Ai-Limi charges for the control plane: seats and routed requests.

What happens if Ai-Limi is down?

Fail-open or fail-closed is a workspace setting. Scale customers can pin a bypass with a signed emergency key.

Can we keep data in the EU?

Team stores metadata in the US. Scale offers an EU-only control plane and regional pinning.