@kuralle-agents/redis-store

Redis-backed session store, memory service, and vector store for Kuralle.

Install

npm install @kuralle-agents/redis-store

Peers: @kuralle-agents/core @kuralle-agents/rag.

What it does

Three backend implementations — sessions, long-term memory, and vector search — all backed by Redis. Works with Upstash, node-redis, ioredis, or any client that exposes compatible get / set / del commands.

Key exports:

• RedisSessionStore — SessionStore implementation for durable session persistence.
• RedisMemoryService — MemoryService implementation for cross-session long-term memory.
• RedisPersistentMemoryStore — PersistentMemoryStore for durable USER/MEMORY markdown blocks.
• RedisVectorStore — VectorStoreCore implementation for vector similarity search.
• fromUpstash / fromNodeRedis / fromIORedis — client adapters.

Session store

import { createRuntime } from '@kuralle-agents/core';
import { RedisSessionStore, fromUpstash } from '@kuralle-agents/redis-store';
import { Redis } from '@upstash/redis';

const sessionStore = fromUpstash(Redis.fromEnv(), { prefix: 'kuralle' });

const runtime = createRuntime({
  agents: [agent],
  defaultAgentId: 'support',
  sessionStore,
});

Client adapters

node-redis:

import { createClient } from 'redis';
import { fromNodeRedis } from '@kuralle-agents/redis-store';

const client = createClient({ url: process.env.REDIS_URL });
await client.connect();
const sessionStore = fromNodeRedis(client, { prefix: 'kuralle' });

ioredis:

import Redis from 'ioredis';
import { fromIORedis } from '@kuralle-agents/redis-store';

const client = new Redis(process.env.REDIS_URL);
const sessionStore = fromIORedis(client, { prefix: 'kuralle' });

Direct constructor (any compatible client):

import { RedisSessionStore } from '@kuralle-agents/redis-store';

const sessionStore = new RedisSessionStore({ client: myClient, prefix: 'kuralle' });

Store options

• prefix (default: 'kuralle') — key namespace.
• sessionTtlSeconds — optional TTL for session keys.
• enableCleanupIndex (default: true) — maintain a sorted set for cleanup by updatedAt.

Long-term memory

import { createRuntime, InMemoryMemoryService } from '@kuralle-agents/core';
import { RedisMemoryService, fromUpstash } from '@kuralle-agents/redis-store';

const redis = fromUpstash(Redis.fromEnv());
const memoryService = new RedisMemoryService({ client: redis });

const runtime = createRuntime({
  agents: [agent],
  defaultAgentId: 'support',
  memoryService,
  preloadMemory: true,
  memoryIngestion: 'onEnd',
});

Working memory blocks

import { createRuntime } from '@kuralle-agents/core';
import { RedisPersistentMemoryStore, fromUpstash } from '@kuralle-agents/redis-store';
import { Redis } from '@upstash/redis';

const client = Redis.fromEnv();
const workingMemoryStore = new RedisPersistentMemoryStore({ client, prefix: 'kuralle' });

const runtime = createRuntime({
  agents: [agent],
  defaultAgentId: 'support',
  defaultWorkingMemoryStore: workingMemoryStore,
});

On Cloudflare Workers, use fromUpstash with the REST client — no TCP socket required.

Related

• @kuralle-agents/core — SessionStore interface and runtime.
• @kuralle-agents/rag — VectorStoreCore interface.
• @kuralle-agents/postgres-store — Postgres alternative.