Skip to main content
Usage Limits let you see where your AI spend is going and cap it — per source and per user — in USD per calendar month. Spend, limits, and caps all reset on the 1st. These controls live in the This month’s spend section of the Credits tab, below the credit balance.

​
By source

Every place Logic AI is called from is a source — the chat itself, plus any integration that calls the global Apex API. Each source row shows its model, spend so far this month, its limit, and whether it’s enabled. Click Edit on a row to:
  • Set a monthly limit — choose No cap or Amount and enter a USD figure. When a source reaches its limit, its calls are refused until the next month (or until you raise it).
  • Change the model the source runs on.
  • Enable or disable the source — a disabled source’s calls are rejected, a quick way to switch off one integration without touching code.
Chat and internal/background callers are grouped together as a single chat bucket; each integration registered through the API appears as its own row (named by the source key the developer chose — see Source Registration).

​
By user

Spend per user, with a monthly Cap you can set per person (No cap or an Amount). A user with no cap of their own falls back to the org default user cap. When a user reaches their cap, their calls are refused until the next month.

​
How limits are enforced

Caps are enforced in addition to your organization’s overall credit balance — a call is refused if it would exceed either an applicable cap or the remaining balance. Buying more credits does not lift a cap, and raising a cap does not add credits; they’re independent controls. When a call is refused for either reason, users see a friendly “out of credits” message in chat, and integrations calling the API get a 402 response — see Errors & Status Codes.