How to Use Redis MCP with Claude Code

What is the Redis MCP?

The Redis MCP connects Claude Code to any Redis-compatible instance — including Redis Stack, Valkey, and cloud-hosted Redis services. Claude can get, set, and delete keys, inspect data structures, analyze memory, and debug caching logic without switching to a separate Redis GUI.

Installation

mcpizy install redis

Provide your Redis connection URL (e.g., redis://localhost:6379 or rediss://user:password@host:6380) during setup. TLS connections are fully supported.

Key Capabilities

• Get, set, and delete keys — read and write any Redis data type (strings, hashes, lists, sets, sorted sets, streams).
• Inspect TTLs and key patterns — scan namespaces and find keys that are expiring unexpectedly.
• Analyze memory usage — identify large keys consuming excessive memory with MEMORY USAGE.
• Debug Lua scripts — test scripts against real data before deploying them.
• Monitor pub/sub channels — publish test messages and verify subscribers receive them.

Example Usage

Your API is returning stale data. Ask Claude to investigate the cache:

// Claude will:
// 1. KEYS user:* — list all user cache keys
// 2. TTL user:12345 — check when the problematic key expires
// 3. GET user:12345 — inspect the cached value
// 4. Identify if the value is outdated and suggest a cache invalidation strategy

Tips and Best Practices

• Avoid using KEYS * on production — use SCAN with a cursor instead to avoid blocking the server.
• Use key namespacing (e.g., session:user:id) so Claude can filter keys logically.
• Combine with your Postgres MCP to compare cached values against the source of truth in your database.
• For Redis Stack users, ask Claude to use RediSearch for complex query patterns on indexed data.

View on the MCPizy directory