How to See Cursor Token Usage by Model and User

Cursor is one of the few AI coding tools that reports usage at the event level — so you can see spend broken down by model and by team member, not just as one monthly number. StackSpend reads Cursor's usage events and rolls them into the AI Explorer alongside your API providers.

Choose your next step

• Checking your own usage? Start in the Cursor dashboard. You do not need another tool to look up your own activity.
• Reviewing a team across several providers? See the Cursor monitoring workflow, then check the team-admin connection requirements.
• Ready to evaluate it? Start a Cursor monitoring trial. The trial goal is to explain one usage change by user and model, verify the available data, and give the follow-up an owner. The free trial lasts 14 days and requires no credit card.

How to check Cursor usage before connecting another tool

Start in your Cursor dashboard. Choose the reporting period and check your team's permissions: admins can see team analytics, while member visibility is more limited. Cursor documents chart exports in its usage analytics guide.

For a team review, record the period, model mix, busiest users and the question you need to resolve. Keep activity measures, tokens and billed spend separate. If you use the API, check event coverage and pagination before comparing totals. An incomplete export is a data gap, not evidence that usage fell.

A useful model-and-user review

1. Compare the same reporting window and time zone in Cursor and your reporting tool.
2. Find the model or user associated with the largest change. Check whether workload, model selection or a repeated task explains it.
3. Review token directions where they are reported; keep request-based events outside token totals.
4. Assign the unexplained change to someone who can check the workload, then repeat the comparison after that investigation.

This is a review workflow, not a productivity ranking. More activity or a larger bill does not establish that an engineer is more or less effective.

What Cursor exposes

Cursor's usage feed reports each event with the model used, the team member who made it, and — for token-based calls — a token breakdown split into input, output, cache-read, and cache-write. That's enough to answer the questions teams actually ask: which models are driving Cursor spend, which developers or teams are heaviest, and how the token mix looks per model.

StackSpend presents the available event data in its Cursor usage monitoring dashboard: Cursor usage by model and by user, with the input/output/cache direction split on token-based calls, and a blended effective rate per model at your real mix.

The one limitation: token-based vs request-based calls

Cursor has two kinds of billed activity, and they carry different data:

• Token-based events can include input, output, cache-read and cache-write counts. Cursor identifies them with isTokenBasedCall and supplies the available counts in tokenUsage.
• Request-based events use a different billing measure. Do not invent token counts for them or treat a token-only total as the whole account's usage.

This is a token-based vs request-based billing distinction, not a StackSpend gap — review the returned events to establish which token directions are available. The AI Explorer's coverage flags make that explicit rather than showing a misleading total.

Cursor's Admin API documentation uses chargedCents for event-cost reconciliation with its team-spend endpoint, including applicable Cursor fees. A token calculation is not a substitute for that billed measure or the invoice. StackSpend's available usage view should be checked against Cursor's own billing records before it is used for reconciliation.

What you can break Cursor spend down by

• Model — which models drive Cursor usage and cost
• User — usage per team member in the connected event feed; project ownership needs a separate verified mapping
• Direction — input / output / cache-read / cache-write, on token-based calls

Connect your Cursor team

You need a Cursor team admin and a Team API key with Admin scope, created under the Team tab rather than as a personal User key. Follow the Cursor connection requirements, then monitor Cursor team spend by user and model. StackSpend reads available usage events; it does not change Cursor account settings.

Initial history depends on what Cursor returns for your team and the requested period. StackSpend requests a backfill and keeps synced events for ongoing trends, but it cannot reconstruct events the API does not return. A full 90-day initial history is not guaranteed.

What to verify during your Cursor trial

Start with one Cursor team and one question: which user or model explains the change in our connected usage costs? After a successful sync, use this checklist before adding more providers.

A useful trial ends with a verified coverage note, one explained change or a documented investigation, and an owner for the next review. It does not require a guaranteed saving. Check plans and feature availability before relying on a budget or alert workflow. StackSpend alerts notify your team; changes to spend limits, licenses or running workloads happen in the relevant provider or workload controls.

Start a Cursor monitoring trial, or review the connection requirements first.

Related

• Cursor team cost monitoring — evaluate the connected user/model view and daily review workflow
• AI Explorer — see how connected model and user breakdowns are presented
• Engineering AI budget worksheet — separate paid seats from additional usage when planning a team budget
• AI spend spike investigation — turn an unexplained change into an investigation with an owner

FAQ

Can I see Cursor token usage by model?

Cursor's team usage events identify the model and user. Token-based events can also include token directions. Check your dashboard permissions, API access and returned event coverage before assuming you have a complete team view.

What is the limitation of Cursor's usage data?

Some calls are billed per request rather than per token, so a pure token view will not reconcile exactly to the invoice. Treat the token breakdown as an attribution tool rather than an accounting one.

How do I attribute Cursor spend to teams?

For the user-level feed connected to StackSpend, map users to the teams responsible for their work. Keep that mapping current and distinguish it from project or repository attribution; a user-level event alone does not establish which customer or project incurred the cost.

References

• Cursor usage analytics: access, filters and exports
• Cursor Admin API: usage-event fields and cost reconciliation
• StackSpend Cursor connection requirements