Available Tools
Redis Operations
"Store the API response in cache with key 'weather:tokyo' and 60 second TTL"
"Get the cached value for key 'user:123:preferences'"
"Increment the rate limit counter for IP 192.168.1.1"
"List all keys matching 'session:*'"Rate Limiting
"Check if user_123 has exceeded 100 requests per minute"
"Set up sliding window rate limiting: 1000 requests per hour per API key"Pub/Sub (Kafka)
"Publish a message to the 'notifications' topic"
"Show recent messages from the 'events' queue"Why Serverless Redis
| Traditional Redis | Upstash |
|---|---|
| Manage server | Zero infrastructure |
| Pay for idle | Pay per request |
| Scale manually | Auto-scales |
| Backup yourself | Built-in persistence |
| TCP connections | REST API (edge-compatible) |
Edge Compatible
Works in Cloudflare Workers, Vercel Edge, Deno Deploy:
import { Redis } from "@upstash/redis";
const redis = Redis.fromEnv();
await redis.set("key", "value", { ex: 60 });Free Tier
- 10,000 commands/day
- 256MB storage
- Global replication
Key Stats
- 2,000+ GitHub stars
- Serverless Redis + Kafka
- Pay-per-request pricing
- Edge-compatible REST API
- Free tier: 10K commands/day
FAQ
Q: What is Upstash MCP? A: An MCP server connecting AI agents to Upstash serverless Redis and Kafka for caching, rate limiting, and messaging with zero infrastructure.
Q: Is Upstash free? A: Free tier: 10K commands/day. Pay-as-you-go after that.
Q: Can I use it for AI agent memory? A: Yes, store conversation context, user preferences, and session data in Redis with TTL for automatic cleanup.