Turn AI from a cost line into a revenue line.
Attribute every token and tool call to the end user who triggered it, apply your margin, and push usage to your billing system. CMD+K becomes a P&L, not an invoice.
Every model token and tool call tagged to the end user and audience.
Set markup per model or per tool; dashboard shows revenue vs cost.
Stripe & Paddle first-class; signed webhook for the rest. No MoR, no Connect — you keep the relationship.
The model
Operators run Command+K with managed AI keys. We meter usage per end user, apply the margin you configure, and forward meter events to your billing system. We never hold end-user funds.
AI Runtime — power your own features too
Beyond the assistant, your backend can call our OpenAI-compatible API to run any AI feature in your product — summaries, generators, agents — on managed model keys, with no provider account. Every call is attributed to the end user you name, shows up on the same per-user dashboards, respects the same plan limits, and forwards to the same billing destination. Prepaid credits, hard stop at zero: your AI spend can never run away.
Guardrails by default
- Per-user cost caps stop runaway usage before it bills.
- Workspace-level spend ceilings as a second safety net.
- Retry pipeline (pgmq) for meter forwarding — no silent drops.
$99/mo to unlock