Understand your AI usage: every agent, model, and request
By Steven Van ·
See what your team spent on every model, save the charts you keep rebuilding, click any bar to land in the logs behind it, and query the same data from…
OpenRouter has added an Activity dashboard and an Analytics API for tracking AI spend down to the individual agent, model and request.
- Overview shows total spend, requests, token volume, cache hit rate and blended cost per million tokens, along with spend by model, top users and apps, and OpenRouter credits versus BYOK spend.
- Trends sorts the same data by movement rather than size, to help spot a runaway agent or a model gaining traction.
- Explore lets you build custom charts by metric, up to two grouping dimensions (model, provider, API key, app, user, workspace and more), and time rollup, then save views for yourself or the whole organization and export them as CSV or PDF.
- Guardrails reports what prompt injection and sensitive-information rules blocked, redacted or flagged, and which rules caught it.
- Any chart or ranked table links through to the underlying logs, down to a Generation detail view (cost, latency, routing, attribution) and a Prompt detail view with a token flamegraph coloured by message role.
The same data is available through a new Analytics API: a metadata endpoint listing supported metrics and dimensions, and a query endpoint for pulling the numbers into a terminal or a custom dashboard. OpenRouter's announcement describes pointing a coding agent at the accompanying openrouter-analytics skill to run a cost review, which traced a preview model burning about $6.2K a month, roughly 25x the org's blended rate, back to a single batch-pipeline key running a task that never needed a frontier model.
