Usage & metering
How Syncanix meters usage with compute units (CU), how they're counted, what happens at your limit, and where to track it.
Syncanix meters usage by what your assistant actually does — not by seats, requests, or servers. The unit is the compute unit, or CU: a weighted measure of messages, tool calls, and the tokens behind them.
What’s a compute unit?
A compute unit (CU) is a weighted measure of the work your assistant does. Each message, each tool or MCP call, and the underlying tokens draw CU — a heavier turn that calls tools or works through a lot of context uses more than a quick question. The meter resets at the start of each calendar month (UTC).
How it’s counted
The meter charges you for what your assistant actually does, weighted by effort. A few rules follow from that:
- Each message and action draws CU by its weight — a heavy user who chats and triggers tools all day costs more than a casual one.
- The meter resets to zero at the start of every calendar month (UTC).
- Your own team using the dashboard isn’t metered — only the work your assistant does for your end users consumes CU.
What a conversation actually costs
CU follow effort, and the weights are simple: a message the assistant answers is 2 CU, a tool or MCP call is 2 CU, an endpoint action is 1 CU, and a generated insight is 5 CU (tokens are folded in on top). So a typical support chat — the assistant answers (2 CU) and calls one tool to look something up (2 CU) — is about 4 CU. Three rough shapes cover most of what you’ll see:
- A quick question — a few CU
- One message in, one answer out, no tools — the assistant just looks something up or explains. This is the lightest kind of turn.
- A multi-step action — more
- A few back-and-forth messages that call one or more of your API actions — say, 'find my open invoice and email it to me.' Each message and tool call adds its weight, so this uses more than a quick question.
- A long, context-heavy session — most
- A drawn-out conversation that pulls in a lot of your knowledge, works through many steps, or calls several tools. Longer context and more tokens mean more CU — the heaviest kind of turn.
Your monthly allowance
Each plan includes a monthly compute-unit allowance; Enterprise is custom. The exact number for each plan is on the pricing page.
Approaching the limit
The dashboard shows your current usage against your allowance, with a status that changes as you get closer:
- Within
- Comfortably under your allowance. Nothing to do.
- Approaching
- You’ve used about 80% of your allowance — a good moment to consider upgrading.
- Reached
- You’ve hit your allowance for the month.
When you reach the cap
Reaching your allowance doesn’t switch your assistant off. By default it pauses new turns once the included compute units are used up, and your data and conversation history stay readable — or you can opt in to pay-as-you-go overage to keep running. The meter clears at the start of the next month, and upgrading raises the allowance immediately.
Tracking usage
Open Settings → Billing in the dashboard to see your current compute-unit usage, your allowance, and where you are in the month. Usage logs break activity down in more detail.