@llm-optimize/cache-redis

Redis-backed LLM response cache for @llm-optimize/core. Share cached responses across multiple servers, workers, or serverless functions.

Install

npm install @llm-optimize/cache-redis @llm-optimize/core

Works with any Redis client that implements get, set, and del — including ioredis and node-redis.

Usage — with ioredis

import { LLMOptimizer } from "@llm-optimize/core";
import { RedisCache } from "@llm-optimize/cache-redis";
import Redis from "ioredis";

const redis = new Redis(process.env.REDIS_URL);

const ai = new LLMOptimizer({
  provider: "openai",
  cache: new RedisCache(redis),
  providerClientFactory: { create: () => myClient },
});

Usage — with node-redis

import { createClient } from "redis";
import { RedisCache } from "@llm-optimize/cache-redis";

const redis = createClient({ url: process.env.REDIS_URL });
await redis.connect();

const cache = new RedisCache(redis);

Custom Key Prefix

// Default prefix is "llm-opt" → keys stored as "llm-opt:<hash>"
const cache = new RedisCache(redis, "my-app");
// Keys stored as "my-app:<hash>"

Cache TTL

LLMOptimizer sets a default TTL of 5 minutes per cached response. You can also use the cache directly:

await cache.set("my-key", { content: "response" }, 120_000); // 2 minute TTL
await cache.get("my-key");
await cache.delete("my-key");

When to Use

| Scenario | Use | |---|---| | Multiple app servers | ✅ RedisCache | | Serverless / Lambda | ✅ RedisCache | | High-traffic production | ✅ RedisCache | | Local dev / single server | Use @llm-optimize/cache-memory instead |