Every scan searches each source through treg, which routes the search to a data provider and charges per call. The Usage tab shows what your project spends.
What a scan is
One scan of one keyword makes one call per source: X, Reddit, LinkedIn, YouTube and the web. Brand keywords are scanned every hour, so each one means five calls an hour; competitors and lead phrases are scanned every 6 hours by default. Under Settings you can set each keyword's interval (every 1, 3, 6, 12 or 24 hours) and turn sources off; a source that is off is not called and costs nothing.
Scan now on the Settings tab runs an extra scan and shows how many mentions it found, how many were new, and what it cost.
The Usage tab
| Number | Means |
|---|---|
| Projected treg cost / month | A month of scans of every keyword at its interval, at the average price per call so far. |
| treg spent · 30 days | What the searches of this project cost in the last 30 days. |
| Calls · 30 days | How many search calls those scans made. |
| Average per call | Spent divided by calls. |
Below them, Spend by provider splits the last 30 days into treg searches, Claude (scoring and classifying posts) and TypeSafe (the lead gate), with calls, tokens and cost, and Spend by keyword shows each keyword's treg and model cost and its cost per day.
Below that is a list of the latest calls: when, which keyword and source, which provider answered, which were tried first, how many mentions it found and how many were new, how long it took and what it cost.
The projection grows with the number of keywords. To spend less, remove keywords that mostly find unrelated posts.
Reading your homepage
Starting a project from a domain reads your homepage once. That call shows in the list too.