Skip to main content
Parlé uses a three-layer budget system to give your firm precise control over AI spend. Project budgets are the innermost layer — they constrain spending within a single engagement. Personal budgets do the same for the personal conversations that belong to no project, and the org-wide cap sits above both as the outer backstop, catching overruns that individual budgets might not anticipate. Understanding how these layers interact helps you tune controls that protect the firm without interrupting day-to-day work.

​
How the Layers Work

Every AI send in Parlé is checked before it is executed: the org-wide cap always, then the project’s caps for a send in a project, or the sender’s personal budget for a send in a personal conversation. If any layer that applies is exhausted, the send is blocked.
  1. Project budget — blocks a specific send in a specific project when that project’s monthly cap or lifetime cap is reached. This is the first and most granular control, scoped to a single client engagement, and it is stated in the project’s own budget currency. See Budgets.
  2. Personal budget — blocks a send in a personal conversation (one that belongs to a person rather than to a project) once that person has used their personal budget for the period, which is a month or a week. Spend in a project counts against the project, never against anyone’s personal budget, so this layer bounds the work that is not charged to any engagement.
  3. Org-wide cap — blocks any send by anyone in the org, in a project or in a personal conversation, once total org-level AI spend reaches the monthly ceiling. This is the ultimate backstop: no single project or user can push the org beyond the limit you set. If your plan carries a monthly ceiling of its own, the lower of the two applies.
Caps apply to model API call costs only — they do not affect your PParlé subscription fees.
The org-wide cap and personal caps are always in US dollars, even if your projects are capped and reported in another currency. The org-wide cap field is labelled Cap (USD / month), personal budgets read with the code on the Members page (for example $25.00 USD / month), and the notice a blocked member sees spells it out too — “Organization monthly spend cap of $10.00 USD would be exceeded.” or “Personal budget of $100.00 USD for this period would be exceeded.” — so the ceiling is not mistaken for your own currency. Only project caps can be set in GBP, EUR or CAD. See Currencies and Exchange Rates.

​
Setting the Org-Wide Cap

1

Open the Budgets & alerts tab

Select Org settings in the sidebar and open the Budgets & alerts tab. The Monthly spend cap card is near the top, just below the Credit wallet.
2

Enter the monthly cap

In Cap (USD / month), enter the maximum US-dollar amount the entire org may spend on AI model calls in a single calendar month, across all projects and personal conversations. If your plan has a ceiling of its own, the card says so — “Your plan limits this to $…/mo” — and the cap cannot be set above it. On a plan with no ceiling, leave the field blank for no cap.
3

Save

Select Save cap. The limit is actiParlémediately. The cap resets automatically on the first of each calendar month, regardless of how much of the period has elapsed.

​
Setting Per-User Caps

A personal budget is how much one person may spend on their own conversations, outside any project. There are two places to set it:
  • The organization default — go to Org settings → Budgets & alerts and find the Personal workspaces card. Enter an amount in Budget per person, choose Each month or Each week, and select Save changes. This applies to everyone who does not have a budget of their own. The same card holds Allow personal conversations: turn it off and members can only work inside projects.
  • One person’s budget — select Members in the sidebar and, in the Personal budget column, select that person’s figure (it reads Default · … until they have their own). Enter the Budget, choose the period, and select Save. To put someone back on the default, open the same dialog and select Use organization default.
A monthly budget resets on the 1st of the month and a weekly one on Monday, at 00:00 UTC.

​
What Happens When a Cap Is Hit

When a send would exceed an active cap, Parlé blocks the request before it reaches the model provider. The member sees an in-app message explaining which limit was reached.
The project has reached its monthly cap or its lifetime cap. The notice names which one, in the project’s budget currency. Anyone on the project can select Request a higher limit; only an Org Admin can approve it or raise the cap directly on the project’s Budget tab. A monthly cap also resets on the 1st of the next month. See Budgets.
The member has used their personal budget for the period, so they cannot send in a personal conversation until it resets. Their project work is unaffected. Only an org admin can raise it: select Members in the sidebar, select the figure in that person’s Personal budget column, and enter a higher Budget — or raise the default for everyone in Org settings → Budgets & alerts, on the Personal workspaces card.
One message was estimated to cost more than the per-message spending limit allows. This is not a budget problem, and the notice says so. Compact the conversation or start a new one; for a project limit, select Request a higher limit. The organization-wide limit is on Org settings → Budgets & alerts, on the Per-message spending limit card.
The entire organization has reached its monthly AI spending ceiling. All sends from all members are blocked until an org admin raises the cap or the month resets. Go to Org settings → Budgets & alerts and raise the Cap (USD / month) on the Monthly spend cap card. If the cap is already at your plan’s ceiling, it can only be raised by upgrading the plan.
Org-wide and personal caps are checked at send time using cost estimates. If a large response causes the final billed cost to slightly exceed a cap, the overage is small and one-time — the next send will be blocked until the limit is raised or the month resets.
For firms new to AI cost management, start with a conservative org-wide cap equal to roughly twice your first month’s expected usage. You can always raise it — and a low ceiling prevents surprise invoices while your team calibrates their workflows.