# Usage & analytics
The portal's **Usage** page shows spend, requests, tokens, latency, and error rate for the
selected workspace. You can filter by model and time range, break it down per model and per key,
and export a CSV. Every API response carries the same numbers (`usage.cost`,
`usage.completion_tokens_details.reasoning_tokens`), so you can track usage from your own logs
without the portal. Keln does not store prompt or completion content
([zero data retention](../capabilities/privacy.md)).
Usage & analytics
The portal's Usage page shows spend, requests, tokens, latency, and error rate for the
selected workspace. You can filter by model and time range, break it down per model and per key,
and export a CSV. Every API response carries the same numbers (usage.cost,
usage.completion_tokens_details.reasoning_tokens), so you can track usage from your own logs
without the portal. Keln does not store prompt or completion content
(zero data retention).