upstash

name	upstash
description	Upstash serverless Redis, rate limiting, QStash job scheduling, and Upstash Vector for semantic search — all edge-compatible
layer	domain
category	database
triggers	["upstash","upstash redis","upstash ratelimit","qstash","upstash vector","serverless redis","edge redis","serverless rate limit"]
inputs	["Use case (caching, rate limiting, job scheduling, vector search)","Runtime environment (Edge, Node.js, serverless)","Throughput and latency requirements"]
outputs	["Upstash client configuration","Rate limiting middleware","QStash job scheduling setup","Vector index and query implementation"]
linksTo	["redis","caching","nextjs","vercel"]
linkedFrom	["redis","caching","authentication"]
preferredNextSkills	["caching","nextjs"]
fallbackSkills	["redis"]
riskLevel	low
memoryReadPolicy	selective
memoryWritePolicy	none
sideEffects	["Creates Upstash resources via REST API"]

Upstash Skill

Purpose

Implement serverless-native Redis, rate limiting, background job scheduling, and vector search using Upstash. All Upstash clients communicate over HTTP/REST, making them compatible with edge runtimes, Vercel serverless functions, and Cloudflare Workers where persistent TCP connections are not available.

@upstash/redis — Serverless Redis

// lib/upstash.ts
import { Redis } from "@upstash/redis";

export const redis = new Redis({
  url: process.env.UPSTASH_REDIS_REST_URL!,
  token: process.env.UPSTASH_REDIS_REST_TOKEN!,
});

// Basic operations — same API as ioredis but over HTTP
await redis.set("user:123", { name: "Jane", plan: "pro" }, { ex: 3600 });
const user = await redis.get<{ name: string; plan: string }>("user:123");

// Pipeline — batches multiple commands into a single HTTP request
const pipe = redis.pipeline();
pipe.incr("page:views");
pipe.expire("page:views", 86400);
pipe.get("page:views");
const results =  pipe.();