mcp-server-vector-memory 0.1.0

MCP server for semantic document memory with Redis vector store

Vector Memory MCP Server

An MCP server that gives AI assistants the ability to save and recall information from files. Works like a long-term memory system where you can store documents and retrieve relevant information later using natural language.

Features

This server exposes two main tools:

1. save_to_memory - Save files to memory for later retrieval

   • Supports PDF, TXT, and MD files
   • Accepts both absolute and relative file paths (converts all to absolute)
   • Automatically processes and organizes content for efficient recall
   • Preserves source file information

2. recall_from_memory - Recall information from memory

   • Find relevant content using natural language queries
   • Returns the most relevant information even without exact word matches
   • Shows absolute paths to source files for each result

Prerequisites

• Python 3.12 or higher
• Redis server running locally on port 6379
• UV package manager

Start Redis

# Using Docker
docker run -d -p 6379:6379 redis:latest

# Or using Homebrew on macOS
brew install redis
brew services start redis

Installation

The dependencies are already configured in pyproject.toml. Install them using:

uv sync

Verify Your Setup

Before running the server, test your configuration:

uv run test_connection.py

This will verify:

• ✅ All required packages are installed
• ✅ Redis is running and accessible
• ✅ Embedding model loads correctly

Usage

Running the Server Standalone

You can test the server directly:

uv run vector_memory.py

Integrating with Claude or Windsurf

Add this configuration to your MCP config file at:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "vector-memory": {
      "command": "uv",
      "args": [
        "--directory",
        "/ABSOLUTE/PATH/TO/mcp-servers/servers/vector-memory",
        "run",
        "vector_memory.py"
      ]
    }
  }
}

Important: Replace /ABSOLUTE/PATH/TO with the full path to your project directory.

After updating the config, restart Claude for Desktop or Windsurf.

Integrating with Codex CLI

For Codex CLI, add this to your MCP configuration file (typically ~/.config/codex/mcp_config.toml or similar):

[servers.vector-memory]
command = "uv"
args = [
  "--directory",
  "/ABSOLUTE/PATH/TO/mcp-servers/servers/vector-memory",
  "run",
  "vector_memory.py"
]

Important: Replace /ABSOLUTE/PATH/TO with the full path to your project directory.

After updating the config, restart or reload Codex CLI.

Tool Examples

Saving to Memory

In Claude/Windsurf/Codex, you can ask:

Save these files to memory:
- /path/to/document1.pdf
- /path/to/notes.md

or simply:

Remember the contents of /path/to/my-notes.txt

Recalling from Memory

What do you remember about machine learning algorithms?

or:

Recall information about project deadlines

Configuration

You can customize the following constants in vector_memory.py:

• REDIS_URL: Redis connection string (default: redis://localhost:6379)
• INDEX_NAME: Vector store index name (default: doc_chunks)
• MODEL_NAME: Embedding model (default: sentence-transformers/all-MiniLM-L6-v2)

Architecture

┌─────────────────┐
│  Claude/Client  │
└────────┬────────┘
         │ MCP Protocol
         │
┌────────▼────────┐
│  Vector Memory  │
│   MCP Server    │
└────────┬────────┘
         │
         ├─────► HuggingFace Embeddings
         │
         └─────► Redis Vector Store

Memory Management

Two scripts are provided to manage stored documents:

Quick Cleanup (cleanup.py)

Simple commands for common operations:

# Show memory statistics
uv run cleanup.py stats

# Delete all documents from memory
uv run cleanup.py all

# Delete documents from a specific file
uv run cleanup.py file /path/to/file.pdf

Advanced Management (manage_memory.py)

Interactive tool with search and selective deletion:

# Interactive mode
uv run manage_memory.py

# Command-line usage
uv run manage_memory.py list                    # List all files
uv run manage_memory.py search <term>           # Search by filename
uv run manage_memory.py delete-file <path>      # Delete specific file
uv run manage_memory.py delete-all              # Delete everything

Interactive Mode Features:

• 📋 List all documents grouped by source file
• 🔍 Search documents by filename or path
• 🗑️ Selectively delete documents with confirmation
• 📊 See chunk counts per file

Development

To run in development mode with auto-reload:

uv run --reload vector_memory.py

Troubleshooting

Redis Connection Error

Ensure Redis is running:

redis-cli ping
# Should return: PONG

Model Download

The first time you run the server, it will download the embedding model (~80MB). This is normal and only happens once.

File Not Found Errors

The server accepts both absolute and relative file paths, but automatically converts them to absolute paths for storage. If a file is not found, check that the path is correct relative to where the server is running.

Path Handling

• Input: Accepts both absolute (/full/path/to/file.txt) and relative (./docs/file.txt) paths
• Storage: All paths are converted to absolute paths before being saved to memory
• Output: recall_from_memory always returns absolute paths to source files

This ensures consistent path references regardless of how files were originally added to memory.