home / mcp / buildautomata memory mcp server

BuildAutomata Memory MCP Server

Memory MCP and CLI

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "brucepro-buildautomata_memory_mcp": {
      "command": "python",
      "args": [
        "C:/path/to/buildautomata_memory_mcp_dev/buildautomata_memory_mcp.py"
      ],
      "env": {
        "BA_USERNAME": "buildautomata_ai_v001",
        "QDRANT_HOST": "localhost",
        "QDRANT_PORT": "6333",
        "MAX_MEMORIES": "10000",
        "BA_AGENT_NAME": "claude_assistant",
        "CACHE_MAXSIZE": "1000",
        "QDRANT_MAX_RETRIES": "3",
        "MAINTENANCE_INTERVAL_HOURS": "24"
      }
    }
  }
}

BuildAutomata Memory MCP Server provides AI agents with a persistent, versioned memory system that can be accessed over the Model Context Protocol. It enables semantic search, memory versioning, and cross-tool memory sharing to give your AI a long-term, context-aware memory store.

How to use

You connect to the BuildAutomata Memory MCP Server through an MCP client. Run the local server, then configure your MCP client to point at the provided stdio MCP entry. Use the server to store new memories, update existing ones to create new versions, search memories semantically, and inspect memory timelines. The tools expose clear actions for storing, updating, searching, timeline retrieval, maintenance, and pruning.

How to install

Prerequisites: Python 3.10+ and an MCP-compatible client (such as Claude Desktop or Claude Code).

# Step 1: Clone the project repository
git clone https://github.com/brucepro/buildautomata_memory_mcp.git
cd buildautomata_memory_mcp-main

# Step 2: Install Python dependencies
pip install mcp qdrant-client sentence-transformers

Configuration and run

Configure the MCP client to connect to the local stdio MCP server using the exact command shown below. This example uses the Windows path from the quick start configuration.

{
  "mcpServers": {
    "buildautomata_memory": {
      "command": "python",
      "args": ["C:/path/to/buildautomata_memory_mcp_dev/buildautomata_memory_mcp.py"]
    }
  }
}

Start the server

Run the command to start the MCP server using Python. The server will create its database on first run.

python C:/path/to/buildautomata_memory_mcp_dev/buildautomata_memory_mcp.py

Available tools

store_memory

Create a new memory entry with optional category, tags, and metadata.

update_memory

Modify an existing memory; creates a new version while preserving history.

search_memories

Perform semantic and full-text search over memories with filters.

get_memory_timeline

Retrieve the complete version history for a memory item.

get_memory_stats

Provide system statistics and memory usage details.

prune_old_memories

Cleanup old or low-importance memories to reclaim space.

run_maintenance

Run database maintenance and optimization tasks.