Platform Credits & Cost

FleetQ meters LLM and agent usage in platform credits. Credits are the unit you spend when runs call models through the platform's AI gateway. They are separate from your subscription plan — the plan sets structural limits and monthly run/send quotas, while credits cover metered model cost.

How credits work

• 1 credit = $0.001 USD. Model cost is computed in USD, then converted to credits.
• Before each run, the platform reserves an estimated credit amount with a 1.5× reservation multiplier (a safety buffer for output variance). After the run completes, the reservation is settled against the actual cost — excess credits are released back, and any shortfall is charged.
• Output is capped at 8192 tokens per call, which bounds the worst-case cost of a single run.
• Reservations use pessimistic locking, so concurrent runs can't overspend a team's balance.

Your usage page

Every team can see its own consumption at /usage. The page shows the team's monthly usage, resource counts, a credit snapshot, and recent failed runs — all resolved from your current team, so there's nothing to configure.

Credit packs & top-ups

When a team needs more credits than its plan includes, it can purchase a credit pack. A purchase adds a fixed number of credits to the team's balance and records the pack key and the charge amount. Partners integrating via the Partner API receive a pack.purchased webhook with the packKey, creditsAdded, and the charge amountMinor (in minor currency units).

Super-admin: cost & credit tooling

These tools are only available to FleetQ super-admins under the /admin area.