# Token analytics — Feature guide | Remote AI

> Understand recorded model usage through trends, heatmaps, and practical breakdowns.

Source: https://remoteai.piperhex.com/en/features/token-usage/

## Who it is for

For people reviewing usage habits, comparing models, or investigating proxy request performance.

## Start with the larger pattern

Heatmaps and trends help you review activity over time and understand changes in usage.

- Review weekly activity heatmaps.
- Follow Token usage trends.
- Relate usage peaks to your work.

## Break usage down by what matters

Compare recorded usage across Token types, Providers, models, and accounts to understand different workflows.

- Distinguish recorded Token categories.
- Compare model and Provider usage.
- Review account-level breakdowns.

## Go from totals to individual requests

Proxy sessions show messages, context usage, and response latency. Export structured diagnostics when you need to investigate further.

- Inspect recorded sessions and messages.
- Review context occupancy and response time.
- Export diagnostic information for troubleshooting.

## Getting started

1. Run tasks through Remote AI GUI or its proxy.
2. Open Token analytics to review trends and breakdowns.
3. Use proxy sessions to investigate individual requests.

## Before you begin

Statistics reflect the requests actually recorded by the application. Cost estimates are not the final bill from your service provider.

### Does this include every model call on my computer?

No. Coverage depends on the requests Remote AI records and the active configuration.

### Is the displayed cost my actual charge?

Amounts are estimates or information returned by a service. Refer to the provider for actual subscription or API billing.

[Read project documentation](https://github.com/piperhex/remoteai/blob/master/README.md)

[GitHub](https://github.com/piperhex/remoteai) · [Apache-2.0](https://github.com/piperhex/remoteai/blob/master/LICENSE)