ai-infra / API Reference

CachingInterceptor

from ai_infra.mcp import CachingInterceptor

ai_infra.mcp

Cache tool call results for a configurable TTL. Caches based on tool name, arguments, and server name. Useful for expensive operations with deterministic results.

Attributes

ttl_seconds: Time-to-live for cache entries (default: 300).

python

cache = CachingInterceptor(ttl_seconds=60)
    mcp = MCPClient([config], interceptors=[cache])

Constructor

CachingInterceptor(ttl_seconds: int = 300, _cache: dict[str, tuple[CallToolResult, float]] = dict()) -> None

Parameter	Type	Default	Description
`ttl_seconds`	`int`	300	—
`_cache`	`dict[str, tuple[CallToolResult, float]]`	dict()	—

CachingInterceptor

Attributes

Methods

CachingInterceptor

Attributes

Methods