memory-journal

Memory Journal MCP Server acts as a persistent memory hub for AI agents working across large projects. It provides a consolidated journal with automatic session handoffs, GitHub integration, and a set of 42 tools organized into groups like core, search, analytics, relationships, backup, and team. The server exposes dual transport options for clients: a modern HTTP endpoint for streaming interactions and a legacy SSE path for compatibility with older clients. With this MCP you can capture and query project context, generate standups or retrospectives from journal entries, and maintain continuity between AI sessions so decisions and learnings aren’t lost between conversations.

Prerequisites:

• Node.js (LTS) installed on your system
• npm (comes with Node.js)

Install (global):

• npm install -g memory-journal-mcp

Run directly (no global install):

• npx -y memory-journal-mcp

Docker alternative (if you prefer containerized deployment):

• docker run -i writenotenow/memory-journal-mcp

Configuration (env vars and options):

• MEMORY_JOURNAL_MCP_TOOL_FILTER: control which tool groups are exposed (default: full). Example: -github to hide GitHub tools.

Examples:

• Run locally with default tool set: npx -y memory-journal-mcp

• Run with a restricted tool filter via environment: MEMORY_JOURNAL_MCP_TOOL_FILTER=-github,note to adjust exposure, then start memory-journal-mcp as usual.

Tips and common issues:

• If you’re deploying in a serverless or containerized environment, consider using the dual HTTP transport and ensure you enable stateless mode if required.
• The tool filter env var is optional but handy for restricting capabilities in production.
• Check the MCP Registry entry for memory-journal-mcp to verify compatibility and available endpoints.
• For backups and restore operations, use the provided journal backup commands within the MCP tool suite.
• Ensure proper IAM and network permissions if you expose the HTTP endpoint publicly.