# Amazon ElastiCache/MemoryDB for Valkey

Amazon ElastiCache/MemoryDB for Valkey is an MCP server for Claude Code with 12 Valkey tools for search indices, JSON paths and tiered command execution.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 9939 repo stars
- Author: awslabs
- License: Apache-2.0 · policy=link-only · description only; the file stays with the author
- Source: https://github.com/awslabs/mcp/tree/HEAD/src/valkey-mcp-server
- Updated: 2026-10-23
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/amazon-elasticachememorydb-for-valkey

## What it does

Twelve tools, grouped into three sets. Index work runs through four of them: one builds, drops and lists indices from a schema you describe, another loads documents and can generate embeddings on the way in, a third handles semantic, keyword, hybrid and similar-item lookups, and the last assembles aggregation pipelines with grouping, sorting and filters.

JSON keys get five tools for reading and writing a value at a path, appending to arrays, popping elements and trimming ranges. The remaining three execute raw commands and are ranked by risk: reads always pass, writes reject destructive commands and switch off under `--readonly`, admin calls stay disabled until `VALKEY_ADMIN_ENABLED=true` is set and `confirm=True` is passed.

Startup is `uvx awslabs.valkey-mcp-server@latest` once `uv` and Python 3.10 are installed, and the datastore needs the Search and JSON modules — both ship in the `valkey/valkey-bundle` image. Embeddings go to Bedrock by default, so with no AWS credentials semantic calls die on `NoCredentialsError`; set `EMBEDDING_PROVIDER` to Ollama or supply `OPENAI_API_KEY` for OpenAI.

## Requirements
- Needs: Docker

## Install

### Claude Code (native platform)

**Install**
- The MCP registry has no launch command for it: see the repository README

## When to use it
- Build a search index over product fields without writing raw search commands
- Load documents into an index with embeddings generated for you
- Patch one field inside a JSON document without rewriting the key
- Aggregate average values grouped by category
- Give an agent read access while destructive commands stay blocked

## Example prompts
- Create a `products` index in Valkey with `title` as TEXT, `category` as TAG and `price` as NUMERIC.
- Load the product documents into the `products` index.
- What is the average price per category in `products`?

## Similar
- [amazon-elasticache](https://affpapa.org/en/nevibecode/skills/amazon-elasticache.md) — skill · 6300 installs
- [amazon-opensearch-service](https://affpapa.org/en/nevibecode/skills/amazon-opensearch-service.md) — skill · 6400 installs
- [valkey](https://affpapa.org/en/nevibecode/skills/valkey.md) — skill · 142 installs
- [swift-valkey](https://affpapa.org/en/nevibecode/skills/swift-valkey.md) — skill · ★ 64 repo stars
- [opensearch-launchpad](https://affpapa.org/en/nevibecode/skills/opensearch-launchpad.md) — skill · 145 installs
- [amazon-polly](https://affpapa.org/en/nevibecode/skills/calesthio-amazon-polly.md) — skill · 134 installs

From the same set: [awslabs/mcp](https://affpapa.org/en/nevibecode/by/awslabs/mcp.md) — 36 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
