Cachebox

The fastest caching Python library written in Rust

---

Cachebox is a high-performance, in-memory caching library for Python. It is written in Rust, has zero Python dependencies, and exposes a familiar dict-like API so you can drop it into existing code with minimal friction.

Key Features

• Extremely Fast

  10–50× faster than other caching libraries — see benchmarks.

• Low Memory Usage

  Roughly half the memory of a standard Python dictionary for equivalent contents.

• Thread-Safe

  All cache operations are protected by internal Rust mutexes.

• Zero Dependencies

  Distributed as pre-built wheels — no Rust toolchain required at install time.

• Full-Featured

  Seven eviction policies, TTL support, @cached decorator, callbacks, and more.

• Compatible

  Python 3.10+ on CPython and PyPy.

When Should I Use Caching?

• Frequent data access — avoid repeated database queries or API calls for the same keys.
• Expensive operations — memoize pure, costly computations so they run only once per input.
• High traffic — absorb load spikes by serving hot data from memory.
• Web page rendering — cache fragments or full pages that are expensive to generate.
• Rate limiting — track counters and windows, or reduce calls to third-party APIs.
• Machine learning — cache predictions for repeated inputs to save inference time.

Quick Example

import cachebox

@cachebox.cached(cachebox.LRUCache(maxsize=128))
def get_user(user_id: int) -> dict:
    # Expensive DB call — cached after the first call
    return db.query("SELECT * FROM users WHERE id = ?", user_id)

# First call hits the database
user = get_user(42)

# Subsequent calls with the same arguments are served from cache
user = get_user(42)

Use a cache class directly when you need full control over keys and lifetime:

from cachebox import FIFOCache

cache = FIFOCache(maxsize=128)
cache["key"] = "value"
assert cache["key"] == "value"
assert cache.get("missing", "default") == "default"

What's Next?

Page	Description
Installation	Install from PyPI with pip or uv
Getting Started	Decorators, key makers, methods, and common patterns
Choosing a Cache	Which algorithm to pick for your workload
Tips & Notes	Pickling, copying, TTL sweepers, stampede prevention
FAQ	Common questions and edge cases
API Reference	Full class and function documentation
Migration Guide	Breaking changes between major versions