Skip to content

🗄️ Backends Overview

FastAPI Cachekit supports multiple cache backends, so you can choose the best fit for your application’s needs—whether you want blazing-fast in-memory caching for development, or distributed caching for production.

Supported Backends

Backend Description Best For Docs
InMemoryBackend Stores cache in the app’s memory (LRU support) Development, testing In-Memory
RedisBackend Uses Redis for distributed, production caching Production, scaling Redis
PostgresBackend Uses PostgreSQL for persistent SQL-based caching Data persistence, SQL Postgres
MemcachedBackend Uses Memcached for high-speed distributed caching High-speed, stateless Memcached
MongoDBBackend Uses MongoDB for document-based distributed caching NoSQL, flexible schema MongoDB
FirestoreBackend Uses Google Firestore for serverless NoSQL caching Serverless, GCP users Firestore
DynamoDBBackend Uses AWS DynamoDB for serverless NoSQL caching Serverless, AWS users DynamoDB
---

How to Choose a Backend

  • InMemoryBackend
  • 🟢 Easiest to set up, no extra dependencies.
  • 🟡 Not shared between processes or servers.

  • 🟡 Data lost on restart.

  • RedisBackend

  • 🟢 Distributed, scalable, and fast.
  • 🟢 Widely used in production.

  • 🟡 Requires a running Redis server.

  • PostgresBackend

  • 🟢 Uses your existing PostgreSQL database.
  • 🟢 Data persists across restarts.

  • 🟡 Slightly slower than in-memory or Redis.

  • MemcachedBackend

  • 🟢 High-speed, distributed, simple.
  • 🟡 No persistence (data lost on restart).

  • 🟡 No built-in authentication by default.

  • MongoDBBackend

  • 🟢 Persistent storage (data survives restarts).
  • 🟢 Built-in TTL index for automatic cache expiration.
  • 🟢 Supports authentication and access control.

  • 🟡 Slower than in-memory caches (e.g., Memcached).

  • FirestoreBackend

  • 🟢 Persistent, serverless NoSQL storage (data survives restarts).
  • 🟢 Native TTL support for automatic cache expiration.
  • 🟢 Scales automatically with Google Cloud infrastructure.
  • 🟢 Supports authentication and fine-grained access control.
  • 🟡 Slightly higher latency compared to in-memory caches.

  • 🟡 Requires Google Cloud project and credentials.

  • DynamoDBBackend

  • 🟢 Persistent, serverless NoSQL storage (data survives restarts).
  • 🟢 Native TTL support for automatic cache expiration.
  • 🟢 Scales automatically with AWS infrastructure.
  • 🟢 Supports IAM authentication and access control.
  • 🟡 Slightly higher latency compared to in-memory caches.
  • 🟡 Requires AWS account and credentials.

Installation for Each Backend

See the Installation Guide for details on installing optional dependencies for each backend.

Backend Setup Guides

Adding More Backends

Want to add support for another backend?
Open an issue or submit a pull request on GitHub!

FastAPI Cachekit makes it easy to switch backends with minimal code changes—just swap the backend class and you’re ready to go!