Skip to main content

​
Overview

MCP (Model Context Protocol) Servers provide API endpoints for external applications and AI models to access your project data securely.

​
What Are MCP Servers?

MCP Servers in MemoryLake are:
  • API Key Endpoints: Authenticated URLs for external access
  • Project-Scoped: Each API key provides access to one project
  • Read-Only: External access cannot modify your data
  • Secure: API secrets shown only once upon creation

​
Key Features

API Key Generation

Create secure API keys for project access with one-time secret display.

​
Use Cases

Provide AI assistants (Claude, ChatGPT, etc.) with read access to your project’s documents and memories.
Build custom applications that query your knowledge base programmatically.
Generate API keys for team members to access project data from their tools.
Create automated workflows that read project data for reporting or analysis.

​
How It Works

1

Generate API Key

Create an API key for your project with optional description.
2

Copy Secret

Important: Copy the API secret immediately - it’s shown only once!
3

Use Endpoint

Use the provided endpoint URL with your API key to access project data.
4

Secure Storage

Store your API secret securely (never commit to version control).

​
What MCP Servers Provide Access To

When you use an API key, external applications can read:
  • Associated documents: File list and contents
  • Memories: All memories in the project
  • Document comments: Comments attached to files
MCP servers provide read-only access. External applications cannot modify, delete, or upload content.

​
Getting Started

Ready to create your first MCP server endpoint?

Create API Keys

Step-by-step guide to generating API keys

​
Next Steps

Creating API Keys

Generate your first API key

API Reference

See API documentation