Elasticsearch Memory

Provides persistent, intelligent memory via Elasticsearch with hierarchical categorization and semantic search for Claude-based workflows.

python

1

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "fredac100-elasticsearch-memory-mcp": {
      "command": "uvx",
      "args": [
        "elasticsearch-memory-mcp"
      ],
      "env": {
        "ELASTICSEARCH_URL": "http://localhost:9200"
      }
    }
  }
}

You can run and use the Elasticsearch Memory MCP Server to give Claude a persistent, intelligent memory store powered by Elasticsearch. It supports hierarchical memory categorization, automatic and batch reviews, and fast semantic search, with safe upgrade paths and efficient context loading.

How to use

Connect your MCP client to the Elasticsearch Memory MCP Server to save memories, load context, and run batch categorization and search workflows. You will save memories with automatic categorization, load a layered initial context (identity, active context, active projects, technical knowledge), search memories with semantic filters, and apply batch reviews to speed up memory organization.

Practical usage patterns you can perform with the MCP server include: saving new memories with content and tags, loading hierarchical context for your current session, reviewing uncategorized memories in batches, applying batch categorizations, and performing semantic searches with category filters.

How to install

Prerequisites you need before installation are Python 3.8 or newer and Elasticsearch 8.0 or newer.

Step 1: Start Elasticsearch. You can run a single-node instance with Docker or install locally.

Install from Python package (recommended for most setups)

Install the MCP package from PyPI.

pip install elasticsearch-memory-mcp

Configure the MCP server for Claude clients

Choose one of the following local runtime options to launch the server and point clients to Elasticsearch.

{
  "mcpServers": {
    "elasticsearch_memory": {
      "command": "uvx",
      "args": ["elasticsearch-memory-mcp"],
      "env": {
        "ELASTICSEARCH_URL": "http://localhost:9200"
      }
    }
  }
}

Alternative: run directly from source

If you prefer contributing or modifying the code, you can run the server from source using a virtual environment.

# Clone the repository
git clone https://github.com/fredac100/elasticsearch-memory-mcp.git
cd elasticsearch-memory-mcp

# Create and activate a virtual environment
python3 -m venv venv
source venv/bin/activate

# Install in development mode
pip install -e .

# Start the MCP server pointing to your ES instance

Claude client configuration examples

If you are using Claude Desktop or Claude Code CLI, you can configure the MCP server to connect to Elasticsearch via the provided commands.

# Claude Desktop (example using uvx)
{"mcpServers": {"elasticsearch_memory": {"command": "uvx", "args": ["elasticsearch-memory-mcp"], "env": {"ELASTICSEARCH_URL": "http://localhost:9200"}}}

What you can do with the MCP server

Available capabilities include saving memories with automatic categorization, loading initial hierarchical context, performing batch reviews of uncategorized memories, applying batch categorization decisions, and performing semantic searches with optional category filters.

Available tools

save_memory

Save a new memory with automatic categorization.

load_initial_context

Load hierarchical context including identity memories, active context, active projects, and technical knowledge.

review_uncategorized_batch

Review uncategorized memories in batches with auto-detected categories and confidence scores.

apply_batch_categorization

Apply categorizations in batch after review to approve, reject, or reclassify memories.

search_memory

Perform semantic search with optional filters by query and category.

auto_categorize_memories

Batch auto-categorize uncategorized memories based on confidence thresholds.