Dev48
Language
  • About
  • Services
  • Industries
  • Technologies
  • Articles
  • Contacts
Book a call
    Home/Articles/Redis cloud remote mcp server now available for your ai agents
Dev48
info@dev48.org+375 29 320-40-96WhatsApp@isky_by

© 2026 · All rights reserved.

Redis Cloud remote MCP server now available for your AI agents

Source: Redis•November 19, 2026 at 11:19 PM

Redis Cloud remote MCP server now available for your AI agents

Your coding agent can already write code against Redis. Managing the Redis Cloud account behind that code still means leaving the chat for the console or the API. The Redis Cloud remote Model Context Protocol (MCP) server closes that gap. From the sa...

Your coding agent can already write code against Redis. Managing the Redis Cloud account behind that code still means leaving the chat for the console or the API.

The Redis Cloud remote Model Context Protocol (MCP) server closes that gap. From the same chat, your AI agent manages your Redis Cloud resources (such as subscriptions, databases, and Agent Memory stores) through a single entry point that Redis hosts at https://cloud.mcp.redis.io/mcp. New Redis Cloud capabilities arrive through the same entry point, with no change on your side.

No local server. No API client to write. No tool list filling up your agent's context.

Here's what it does and how to connect.

How it works

Redis runs the Redis Cloud remote MCP server for you and your agent just connects over standard MCP. The server speaks the protocol over Streamable HTTP at one endpoint, so there's no SDK or plugin. Any MCP client that can send custom request headers works.

Each request goes through three steps:

  • Your client calls the endpoint. It sends your API account key and API user key as headers on every request. With API keys, there's no OAuth flow to script and no token to refresh.
  • The server checks the keys and routes the call. Each Redis Cloud product area, such as databases or Agent Memory, has its own tools. The server puts them behind one endpoint, one set of keys, and one tool catalog, then passes each call to the tool that owns it.
  • The tool calls Redis Cloud with your own credentials. The server has no Redis Cloud credentials of its own, so every call runs with your keys and your permissions.

The server uses standard MCP features: tools for actions, elicitation for confirmations, and MCP Apps for interactive views in clients that support them. Tools return structured results with fixed fields, so your agent reads values instead of parsing free text.

The server manages the resources in your Redis Cloud account. To read or write the data inside a database, use Redis MCP, which connects to a database directly.

It also works with the Redis Docs MCP server at redis.io/mcp. The Redis Docs server tells your agent how Redis works and the Redis Cloud remote server acts on your Redis Cloud account.

If you'd rather run a server yourself, Redis also publishes a local Redis Cloud MCP server.

Connect your client

The quickest way to set up is from the agents page in the Redis Cloud console. Pick your agent, and the page shows the configuration to copy and links to where you create your API keys.

To set it up by hand, you need a Redis Cloud API account key and an API user key. Every client uses the same three settings:

In Claude Code, add the server with one command:

In Codex, set the REDIS_CLOUD_API_KEY and REDIS_CLOUD_SECRET_KEY environment variables to your API account key and API user key, then add the server to ~/.codex/config.toml:

Any other client works if it supports Streamable HTTP and custom request headers. The docs have set up steps for more clients. To check the connection, ask your agent to list your Redis Cloud subscriptions.

What you can ask

You ask in your own words, and your agent finds the right Redis Cloud tool. For example:

  • "List my subscriptions and their databases."
  • "Create a free database on AWS in us-east-1."
  • "Which regions can I use on Google Cloud?"
  • "Show the metrics for my Agent Memory store."

Keep your agent’s context small

The server uses progressive disclosure, so it doesn't fill up your agent's context. Your agent sees only a small, fixed set of entry tools. When it needs to do something, it searches for the matching Redis Cloud tool, gets that tool's full definition, and then calls it. The definitions of every other tool don’t load into your agent’s context, no matter how many tools the server has.

The same design means new tools show up without any change to your client setup. The docs list what the server can do today.

Safe by default

The server can do only what your API user key can do in Redis Cloud. A key for a read-only user can only read, the key's CIDR allow list still applies, and payment card details are masked.

Changes that cost money or are hard to undo ask for your confirmation in your client, through MCP elicitation. To delete a resource, you type its name.

If you decline or cancel, nothing changes, and the server never retries a change by itself. If a change times out, check the resource before you try again.

Get started

Open the agents page in the Redis Cloud console. It walks you through every step: creating an API key, saving it in your environment, adding the server to your agent, and trying your first request.

The full setup, limits, and troubleshooting are in the Redis Cloud remote MCP server docs.

← All articles

What this article says

More in Software Development

All →
¿Quién es Matthew Pittinsky, el nuevo CEO de Blackboard?
Blackboard

¿Quién es Matthew Pittinsky, el nuevo CEO de Blackboard?

How to read an OCR benchmark before you trust it
Datalab

How to read an OCR benchmark before you trust it

Ignite 2026: Building the frontier ecosystem
Microsoft

Ignite 2026: Building the frontier ecosystem

AI Will Influence $75.8 Billion in Cyber Week 2026 Sales: What Online Sellers Should Know
Ordoro

AI Will Influence $75.8 Billion in Cyber Week 2026 Sales: What Online Sellers Should Know

Amazon Black Friday 2026 Runs 12 Days: What Sellers Need to Know
Ordoro

Amazon Black Friday 2026 Runs 12 Days: What Sellers Need to Know

Elastic and HoneyHive join forces to advance AI observability and evaluation
Elastic

Elastic and HoneyHive join forces to advance AI observability and evaluation

More from Redis

Stop repeating yourself: integrate Azure Managed Redis with your databases
Redis

Stop repeating yourself: integrate Azure Managed Redis with your databases

Run RediSearch, vector sets, & RedisJSON natively on Windows with Memurai
Redis

Run RediSearch, vector sets, & RedisJSON natively on Windows with Memurai

Give your Vercel AI SDK agents memory with Redis Agent Memory
Redis

Give your Vercel AI SDK agents memory with Redis Agent Memory

Introducing the joint reference architecture for enterprise-grade Redis on VMware Cloud Foundation 9.1
Redis

Introducing the joint reference architecture for enterprise-grade Redis on VMware Cloud Foundation 9.1

The next chapter of Redis Iris: what agents need to work in the real world
Redis

The next chapter of Redis Iris: what agents need to work in the real world

Agents Are Software: Introducing the Next Chapter of Redis Iris
Redis

Agents Are Software: Introducing the Next Chapter of Redis Iris