AI Cost Optimisation

The Optimisation engine finds ways to cut your AI cost without changing what your app does. It analyses your real traffic and surfaces two kinds of savings:

Model benchmarking

Swap to a cheaper model that performs as well as the one you run today.

Prompt caching

Optimise the capture the reuse of identical prompt prefixes across calls that your provider prices cheaper.

⚠️ Optimisations replay your prompt content, which requires capture message content in the auto-instrument step to be set to True (default).