Skip to main content
Basic Memory exposes a comprehensive set of Model Context Protocol (MCP) tools that enable AI assistants to interact with your knowledge base. These tools are organized into the following categories:

Content Management

Tools for creating, reading, updating, and deleting markdown notes:
  • write_note - Create markdown notes with semantic observations and relations
  • read_note - Read notes by title, permalink, or memory:// URL
  • edit_note - Edit notes incrementally with various operations
  • delete_note - Delete notes or directories
  • move_note - Move notes to new locations
  • read_content - Read raw file content (text, images, binaries)
  • view_note - View notes as formatted artifacts

Knowledge Graph Navigation

Tools for navigating and exploring the knowledge graph:

Search & Discovery

Powerful search capabilities across your knowledge base:
  • search_notes - Full-text search with advanced filtering
  • search - OpenAI-compatible search interface

Project Management

Tools for managing Basic Memory projects:

Visualization

Tools for creating visual representations:
  • canvas - Generate Obsidian canvas files for knowledge graph visualization

Project Resolution

All MCP tools use a unified project resolution hierarchy:
  1. Single Project Mode - Server constrained to one project (via BASIC_MEMORY_MCP_PROJECT environment variable)
  2. Explicit project parameter - Specify which project to use in the tool call
  3. Default project - Server configured default if no project specified
When the project is unknown, use list_memory_projects() or recent_activity() to discover available projects.

Cloud Mode

Basic Memory supports both local and cloud operation modes:
  • Local Mode - Direct access to local markdown files
  • Cloud Mode - Synced access via Basic Memory Cloud
  • Hybrid Mode - Different projects can be local or cloud independently
Projects are automatically routed to the appropriate backend based on their configuration.

Authentication

Cloud projects use JWT-based authentication with subscription validation. Authentication is handled automatically by the MCP server based on the project configuration.

Output Formats

Most tools support both text and json output formats:
  • text - Human-readable markdown formatted output (default)
  • json - Machine-readable structured data for programmatic use
Specify the output format using the output_format parameter.

Build docs developers (and LLMs) love