# CLAUDE.md Guidelines
## Project Structure
This is a memory server for MCP (Multi-Client Protocol) that uses PostgreSQL with pgvector for vector similarity search.
## Commands
- `npm start`: Start server with Node.js
- `npm run dev`: Start with nodemon for development (auto-reload)
- `npm run prisma:migrate`: Initialize database (mentioned in README but not in package.json)
## Code Style
- Use ES Module imports (type: "module")
- camelCase for variables and functions
- Descriptive naming with clear purpose
- Structured error handling with try/catch and detailed logging
- JSON-RPC response format with proper error codes
## Database
- PostgreSQL with pgvector extension
- Use UUID primary keys with gen_random_uuid()
- Store JSON content with JSONB type
- Include created_at/updated_at timestamps
## Best Practices
- Use sendLogMessage() for consistent logging with levels
- Document functions with clear comments
- Include parameter descriptions in method definitions
- Validate inputs before use
- Use environment variables via dotenv for configuration
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/sdimitrov/mcp-memory'
If you have feedback or need assistance with the MCP directory API, please join our Discord server