#See what your processes cost in slots

The Billing page lists every executable process and every agent with the slots per month it would cost under the old and the new license agreement, computed from its currently saved definition.

#Open the Billing page

In the Designer, open your user menu at the bottom left of the sidebar and choose Billing.

flowchart LR
  M["User menu<br/>(bottom left)"] --> B["Billing"]
  B --> T["License agreement toggle:<br/>Old / New"]
  B --> S["Totals per agreement"]
  B --> L["One row per process and agent:<br/>tier · size · scope · slots"]

The page shows:

• one row per executable process and per agent, with its tier, size, scope (elements or connections) and slots per month;
• the total for each license agreement side by side, with the selected one highlighted and the difference between the two;
• a toggle between Old license agreement and New license agreement (old is selected when the page opens);
• a filter to show only processes or only agents.

Select a name to open that process or agent.

A process marked as documentation only (not executable) is not listed: it never runs, so it costs no slots.

⚠️ The values are an estimate of what each definition would cost. They are computed from the definition as it is currently saved — whether it is deployed, and to which environment, is not taken into account, and nothing on the page is an invoice. Unsaved changes in an open editor are not included.

#How a process is sized

The tier is decided by the most demanding element in the process; the size by the number of elements.

What counts:

• Every step of the model counts as an element, including the steps inside an embedded sub-process. Sequence flows never count.
• Start and end events count under the new agreement. Under the old one, only start and end events with a trigger (timer, message, …) count; plain ones do not.
• A call activity counts as one element. The process or agent it calls is listed and priced on its own row.
• A tool modelled inside an agent counts as one element, however many steps it has. It is part of the agent and has no row of its own.
• Collaborate: user tasks, manual tasks and call activities. Under the old agreement, an embedded sub-process or transaction also makes a process Collaborate; under the new one it does not.
• Intelligent (old agreement only): the process calls an agent, or is an agent. Under the new agreement, calling an agent keeps the process at Collaborate.
• A service task that calls a language model is a service task — Automate under both agreements.
• ⚠️ A process with more than 50 elements should be split into sub-processes. The page still prices it as L and marks the row.

#How an agent is sized

Under the new license agreement an agent is priced by its connections, not by its elements:

flowchart LR
  SE["Each start event<br/>(chat, task, webhook)"] --> C(("Connections"))
  AA["Each selected<br/>adapter action"] --> C
  MCP["Each enabled<br/>MCP server"] --> C
  TL["Each tool in<br/>the toolbox"] --> C
  KV["Each knowledge base"] --> C
  C --> SZ["S up to 6 → 4 slots<br/>M 7–14 → 8 slots<br/>L 15–30 → 10 slots"]

• An MCP server counts once, however many of its tools are enabled. A disabled server does not count.
• Every tool in the toolbox counts once, whatever kind it is. A tool that is a chain of steps is one tool.
• An agent costs at least 4 slots.
• ⚠️ Above 30 connections the agent is priced as L and the row is marked.

The row shows how the connections add up, for example "2 start · 3 actions · 1 MCP · 3 tools · 2 knowledge bases".

Under the old license agreement an agent is sized like a process — by its elements — in the Intelligent tier. Each tool in its toolbox counts as one element, however many steps the tool has.

#Rows without a price

A process or agent that has never been saved shows No definition saved; one whose definition cannot be read shows Definition not readable. Both count as 0 slots in the totals — open and save the process to price it.