AnthropicCacheAdapter

Wraps anthropic.Anthropic and anthropic.AsyncAnthropic messages.create calls with response caching. Returns cached results for identical (model, messages, system, params) combinations.

Installation

pip install 'omnicache-ai[anthropic]'

---

Usage

Sync

import anthropic
from omnicache_ai import CacheManager, InMemoryBackend, CacheKeyBuilder
from omnicache_ai.adapters.anthropic_adapter import AnthropicCacheAdapter

client = anthropic.Anthropic()
manager = CacheManager(
    backend=InMemoryBackend(),
    key_builder=CacheKeyBuilder(namespace="myapp"),
)
adapter = AnthropicCacheAdapter(client, manager)

response = adapter.messages_create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Explain vector embeddings"}],
)

# Second call — cache hit, no API cost
response = adapter.messages_create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Explain vector embeddings"}],
)

Async

client = anthropic.AsyncAnthropic()
adapter = AnthropicCacheAdapter(client, manager)

response = await adapter.amessages_create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Explain vector embeddings"}],
)

With system prompt

response = adapter.messages_create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system="You are a concise technical writer.",
    messages=[{"role": "user", "content": "What is RAG?"}],
)

Combine with PromptCacheLayer

For long system prompts, use PromptCacheLayer alongside AnthropicCacheAdapter:

• AnthropicCacheAdapter caches the full response (eliminates entire API calls)
• PromptCacheLayer injects cache_control on system prompts (saves tokens within live calls)

Invalidate by model

adapter.invalidate_model("claude-sonnet-4-6")

---

API Reference

Method	Description
messages_create(**kwargs)	Cached client.messages.create() (sync)
amessages_create(**kwargs)	Cached client.messages.create() (async)
invalidate_model(model)	Invalidate all cached responses for a model