Cache LLM responses to reduce costs and improve speed.
Caching is enabled by default:
Default: .testagent_cache/
.testagent_cache/
Output:
Direct cache access:
| Field | Type | Description |
|-------|------|-------------|
| output | str | Tested output |
output
str
| criteria | str | Evaluation criteria |
criteria
| expected | str | Expected output |
expected
| model | str | LLM model used |
model
!!! tip "When to Cache"
!!! warning "When Not to Cache"