Features / Per-customer budgets
Per-customer LLM budgets, so one customer cannot sink the rest.
If you sell a product built on AI, your customers share your OpenAI or Anthropic account. One customer's stuck agent or a malicious prompt can use up the budget and the rate limit for everyone. ACPpotamia gives each of your customers their own monthly budget and requests-a-minute limit, checked before every call.
How it works.
One header
Send X-ControlPlane-Tenant-Id with each call to say which customer it is for. New customers appear automatically.
A default and exceptions
Set a default for every customer, then give larger customers their own limit. A budget of 0 blocks a customer at once.
Isolated, even under load
Each customer's budget is reserved before the call is sent, so many calls at once cannot overspend it, and a blocked customer does not slow anyone else.
Automate it
Set and read limits from your own code with the Tenant API, for example when a customer signs up or changes plan.
Per API key and per task too
Cap one key's monthly spend, or limit each task's requests a minute and total spend.
Set a customer's budget when they sign up
curl -X PUT https://acp.examopotamia.com/v1/controlplane/tenants/customer_4821 \
-H "Authorization: Bearer acp_your_admin_key" \
-d '{"monthlyBudgetUsd": 25, "requestsPerMinute": 60}'
Questions.
What is a tenant?
One of your own customers. You choose the id, for example your internal customer id, and send it in X-ControlPlane-Tenant-Id.
What does a customer see when they reach their budget?
Their calls get HTTP 402 with the code tenant_budget_exhausted, so your product can show its own message or offer an upgrade. Other customers are not affected.
Which plans include it?
Per-customer budgets are part of Pro and Enterprise. Per-key and per-session limits are on every plan.
Can I see what each customer spends?
Yes. The Tenants page and the Tenant API show each customer's spend, requests and tokens for the month.
Related features.
Budgets and cost control
Put a hard monthly budget on OpenAI and Anthropic spend. ACPpotamia reserves each AI agent call's cost before it is sent, so agents cannot overspend.
Rate limiting
Rate limit every AI agent's calls to OpenAI and Anthropic per team, at the edge. Requests over the limit get HTTP 429 with Retry-After before they cost money.
Loop detection
Detect and stop AI agents stuck in loops: repeated tool calls, retries and identical prompts. ACPpotamia stops the agent in seconds, before it burns your budget.