credentum/veris-memory-mcp-server

Veris Memory MCP Server

A Model Context Protocol (MCP) server that exposes Veris Memory capabilities to Claude CLI and other MCP-compatible hosts.

Overview

This server implements the Model Context Protocol specification to provide Claude CLI with direct access to Veris Memory's context storage and retrieval capabilities. It acts as a bridge between Claude and the Veris Memory backend API.

[Claude CLI] ←→ [Veris Memory MCP Server] ←→ [Veris Memory API]
     |                     |                        |
   JSON-RPC            MCP Protocol            veris-memory-mcp-sdk
   over stdio          Implementation             (Client SDK)

Features

• 🔌 Claude CLI Integration: Native MCP protocol support for seamless Claude CLI integration
• 📦 Complete Tool Set: Store, retrieve, search, and manage contexts in Veris Memory
• 🚀 High Performance: Async implementation with connection pooling and caching
• 🔒 Secure: Proper authentication, input validation, and error handling
• ⚙️ Configurable: Flexible configuration system supporting various deployment scenarios
• 📊 Observable: Comprehensive logging and monitoring capabilities

Available Tools

Quick Start

Installation

# Install from PyPI (coming soon)
pip install veris-memory-mcp-server

# Or install from source
git clone https://github.com/credentum/veris-memory-mcp-server.git
cd veris-memory-mcp-server
pip install -e .

Configuration

1. Set up environment variables:

   export VERIS_MEMORY_API_KEY="your-api-key"
   export VERIS_MEMORY_USER_ID="your-user-id"
   

2. Add to Claude CLI:

   claude mcp add veris-memory \
     --env VERIS_MEMORY_API_KEY \
     --env VERIS_MEMORY_USER_ID \
     -- veris-memory-mcp-server
   

3. Test the integration:

   # Claude CLI will now have access to Veris Memory tools
   # You can use them in conversations like:
   # "Store this decision in Veris Memory: We chose React for the frontend"
   # "Retrieve contexts about our API architecture decisions"
   

Usage Examples

Storing Context with Claude

User: Store this architectural decision: We've decided to use PostgreSQL 
      as our primary database because of its ACID compliance and JSON support.

Claude: I'll store that architectural decision for you.

[Claude uses store_context tool]
✅ Stored context successfully with ID: arch_001

Retrieving Context with Claude

User: What database decisions have we made previously?

Claude: Let me search for database-related decisions.

[Claude uses retrieve_context tool with query "database decisions"]

Found 3 relevant contexts:
1. PostgreSQL selection (arch_001) - Primary database choice
2. Redis caching (arch_005) - Caching layer decision  
3. MongoDB migration (arch_012) - Document store evaluation

Configuration

Configuration File

Create a config.json file:

{
  "veris_memory": {
    "api_url": "https://api.verismemory.com",
    "api_key": "${VERIS_MEMORY_API_KEY}",
    "user_id": "${VERIS_MEMORY_USER_ID}",
    "timeout_ms": 30000
  },
  "server": {
    "log_level": "INFO",
    "max_concurrent_requests": 10,
    "cache_enabled": true,
    "cache_ttl_seconds": 300
  },
  "tools": {
    "store_context": {
      "enabled": true,
      "max_content_size": 1048576,
      "allowed_context_types": ["*"]
    },
    "retrieve_context": {
      "enabled": true,
      "max_results": 100,
      "default_limit": 10
    }
  }
}

Environment Variables

Development

Setup Development Environment

# Clone repository
git clone https://github.com/credentum/veris-memory-mcp-server.git
cd veris-memory-mcp-server

# Install development dependencies
pip install -e ".[dev]"

# Install pre-commit hooks
pre-commit install

Running Tests

# Run all tests
pytest

# Run with coverage
pytest --cov=src/veris_memory_mcp_server --cov-report=html

# Run specific test types
pytest -m unit          # Unit tests only
pytest -m integration   # Integration tests only

Code Quality

# Format code
black src tests
isort src tests

# Type checking
mypy src

# Linting
flake8 src tests

# Run all quality checks
pre-commit run --all-files

Architecture

Core Components

• Server: Main MCP server implementation with JSON-RPC handling
• Protocol: MCP protocol handlers and stdio transport
• Tools: Individual tool implementations (store, retrieve, search, etc.)
• Client: Wrapper around veris-memory-mcp-sdk for backend communication
• Config: Configuration management and validation

Tool Implementation

Each tool follows a consistent pattern:

from veris_memory_mcp_server.tools.base import BaseTool

class StoreContextTool(BaseTool):
    name = "store_context"
    description = "Store context data in Veris Memory"
    
    async def execute(self, arguments: dict) -> dict:
        # Validate input
        # Call Veris Memory API
        # Return result

Troubleshooting

Common Issues

Server fails to start

• Check that all required environment variables are set
• Verify Veris Memory API credentials
• Ensure Python 3.10+ is installed

Tool calls timeout

• Increase timeout in configuration
• Check network connectivity to Veris Memory API
• Verify API key has proper permissions

Claude CLI doesn't see tools

• Restart Claude CLI after adding the MCP server
• Check server logs for initialization errors
• Verify MCP server is configured correctly

Debug Mode

Enable detailed logging:

export VERIS_MCP_LOG_LEVEL=DEBUG
veris-memory-mcp-server

Contributing

We welcome contributions! Please see our for details.

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the file for details.

Support

• Documentation: GitHub Wiki
• Issues: GitHub Issues
• Discussions: GitHub Discussions

Related Projects

• veris-memory-mcp-sdk - Python SDK for Veris Memory
• Model Context Protocol - Official MCP specification
• Claude CLI - Claude Code command line interface

Built with ❤️ by the Veris Memory Team