Claude tokens
0No model dataLocal AI activity
See what Claude Code and Codex used on this computer.
Waiting for the first usage snapshot.
How to read the dashboard
Two different kinds of numbers
Codex tokens
0No model dataCombined tokens
0Claude and Codex togetherAPI-equivalent value
$0.00Planning estimate · not billedUsage trend
Tokens by agent
Model detail
Current model breakdown
| Agent | Model | Input | Output | Cache read | Cache create | Total | Est. cost |
|---|
ChatGPT subscription on the VPS
Checking the subscription connection.
Subscription status is unavailable.
Refresh to check the VPS connection again.
Plan
—Checking connectionCurrent model
—Configured on the VPSCurrent window
—Waiting for usage limitsLonger window
—Waiting for usage limitsShared subscription usage
These limits cover the connected ChatGPT subscription across its apps. The Overview and Usage Planning sections show collected token activity and estimated API-equivalent value; they are not subscription charges.
Usage windows appear when the subscription service reports them.
Receptionist API usage
Activity on the dedicated NOVE Assistant key, funded by the included ChatGPT Pro API-credit benefit.
API consumption
Checking available usage records.
- Text requests
- —
- Text tokens
- —
- Observed voice checks
- —
- Voice duration
- —
- Calculated API cost
- —
- OpenAI posted cost
- Unavailable
Calculated from recorded usage and published OpenAI rates, before included API credits.
Included API credits
Shared across the linked Personal API organization.
- Last verified available
- —
- Current live balance
- Unavailable
- Current grant expires
- —
- Expected next grant
- —
Loading the dated billing snapshot.
How the included credits workUsage breakdown
Final provider usage only; cached tokens are included in input totals.
| Period | Model / activity | Requests | Uncached input | Cached reads | Cache writes | Output | Voice sec | API cost · USD |
|---|
Current rates used in the calculation
| Model | Uncached / 1M | Cached / 1M | 30m write / 1M | 12h write / 1M | Output / 1M |
|---|
Reporting source
Checking sources.
Assistant usage stays separate from local Claude/Codex activity and the Codex subscription windows. Included credits fund eligible API usage; they do not add to those subscription allowances.
Posted cost history
Provider costs before included credits, in USD. These postings do not identify credit deductions or payment-method charges.
| UTC date | Usage item | Posted cost |
|---|
Local usage planning
API-equivalent future scenario
This estimates what the selected Claude and Codex workload could cost at comparable API rates. It does not predict your OpenAI bill.
Waiting for local usage data.
Collector status
See when this dashboard last received fresh local usage totals.
Checking the local data collector.
Successful data updates
One update shown per day
An update is a saved aggregate report—not a separate AI session, charge, or invoice. Extra setup and test runs from the same day are combined here.