Hub
Usage and cost
Hub > Usage shows the tokens every Jennefer app used on this computer and on your endpoint, with cache reads and writes and an estimated cost at list prices. Learn where cost shows in each app.
On this page
Hub > Usage shows how many tokens your Jennefer apps used and, for your endpoint, what that would cost at list prices. Only numbers are kept, never what you wrote or what the model answered.
The three tabs#
| Tab | What it shows |
|---|---|
| This computer | Tokens of the models on this computer, and what the same work would cost on a cloud model you pick. |
| Your endpoint | Tokens and the estimated cost of requests to your endpoint, with cache reads and writes. |
| Jennefer Cloud | Your credits. The tab says Not open yet: Jennefer Cloud models are not available. |
Pick the period at the top right: Days (last 30 days), Weeks (last 12 weeks) or Months (last 12 months).
This computer#
- Tokens on this computer: the total, split into input and output, with the number of requests.
- The same work in the cloud: pick a cloud model, and the card shows what your tokens would cost at its list prices. The model on this computer has no per-token bill.
- Over time: a chart by day, week or month.
- By app and By model: where the tokens went.
- Cloud prices: the price per 1M input and output tokens of each model in the table. Hover a model for its source and notes.
Your endpoint#
- Your endpoint with the period: the estimated total, marked with ~, the number of requests and how much of the input was read from the cache.
- The token split: input, cache read, cache write and output, for all requests and for each model.
- A model without a price in the table shows no price and is left out of the total.
- Over time and By app, with the estimated cost of each app.
Each request's cost is worked out when it ends, with that day's prices, so older numbers do not change when prices change. Claude on Amazon Bedrock outside the global. route is counted 10% higher, like Bedrock's own regional prices.
What the tokens mean#
| Part | Meaning |
|---|---|
| Input | What the model read. On the Your endpoint tab, the input that did not come from the cache. |
| Cache read | Input read from the provider's prompt cache. On Claude it costs about a tenth of normal input. |
| Cache write | Input written to the cache for the next requests. On Claude it costs a quarter more than normal input. |
| Output | What the model wrote, its thinking included. |
Prompt caching for Claude is explained in Your endpoint.
The price table#
Costs use each provider's public list prices, from a table that comes with Jennefer and is updated with its releases. The note under the numbers gives the table's date, for example as of October 21, 2026.
Cost in the apps#
A dollar amount shows only for endpoint models, always as an estimate with ~. Models on this computer show tokens only.
| App | Where |
|---|---|
| IDE | The model chip shows an endpoint model in blue with this window's estimated cost. Its panel, This session, splits the tokens and shows each agent. |
| CLI | The footer shows the cost in blue, for example ~$0.14. /cost prints the conversation's tokens, cache use and cost per model and per agent. |
| CLI+ | The HUD shows the endpoint model, its effort and the conversation's cost. /cost has the details. |
| Workshop | Each build ends with its tokens and, on your endpoint, its cost, for example 42K tokens · ~$0.09. Auto-fix rounds count toward their build. |
| Mobile | A chat's header shows its total tokens and estimated cost, and each answer shows its model, tokens and cost. Build by chat shows the same for each build. |
What is counted#
- Text requests from every Jennefer app, your connected apps and your phone when it uses this computer.
- Images and video are not counted.
- A request stopped before the answer ended is estimated, and the page says how many were.
- Phone chats and builds on an endpoint stored in your account run on Jennefer's servers, not on this computer. They are not in Hub > Usage; the phone shows their tokens and cost.
- The numbers never leave this computer. Clear history deletes them after a second click on Click again to clear.
FAQ#
Is this my bill?
No. It is an estimate from list prices. Your provider's bill is the real number: taxes, discounts and regional prices can differ.
Why does a model show no price?
Its model ID is not in the price table. Its tokens are counted, but its cost is left out of the total.
Is my usage sent to Jennefer?
No. The counts stay on this computer and are not part of usage data sharing.
What does the Jennefer Cloud tab show?
Only your credits balance. Jennefer Cloud models are not available, so the tab has no usage to show.