Credits and usage
Credits are the unit of LLM usage. Every message the agent generates, every tool call it reasons about, every background workflow run draws against your monthly credit balance.
Where to see usage — Settings → Team Settings → Billing. The page shows:
- Your current plan and how many projects it includes
- The Credit Usage bar — how much of your monthly included credits you've spent so far
- A Usage Breakdown section listing spend by model and by project
- (For admins) Overage Settings
Spend by model — inside Usage Breakdown, every model you've used is listed with request count and cost. Expensive models (Opus, GPT-5.4, Gemini 3.1 Pro) rack up credits faster than cheap ones.
Spend by project — if your team has multiple projects, you'll see per-project totals so you can see which project is driving usage.
Overage settings — if the Allow overage charges toggle is off, the agent stops working once your included credits are exhausted. If the toggle is on, the agent keeps working past the included amount and the team is billed for the overage.
Overage limit — when overage is allowed, you can optionally set a maximum overage amount (in dollars). Leave it blank for unlimited. Set it to, say, 50 to cap additional spend at $50/mo beyond included credits.
BYOK and credits — if you have a BYOK key set for the model you're using, the request doesn't hit Magister's credit balance at all. You pay the provider directly and your usage shows as $0 in the Magister breakdown.
Who can change these — only team owners and team admins can change the overage toggle and limit. Members can see the usage page but not edit it.