mem/ctl

mem/ctl is a cloud MCP server that gives AI coding agents persistent memory across sessions. Your agent remembers coding conventions, architecture decisions, lessons learned, and project context - stored server-side and available from any machine.

The problem

AI agents lose context between sessions. Every new conversation starts from zero - your agent forgets your team's coding style, past decisions, and hard-won lessons.

How mem/ctl solves it

1. Authenticate once - run npx memctl auth to store your API token locally
2. Add to your MCP config - just org and project, no token in project files
3. Your agent remembers - context persists across sessions, machines, and IDEs

What you get

• 11 MCP tools with 90+ actions for memory management, context assembly, session tracking, branching, and more
• 9 built-in context types - coding style, architecture, testing, constraints, lessons learned, workflow, folder structure, file map, branch plans
• Cross-IDE - works with Claude Code, Cursor, Windsurf, VS Code, Cline, Roo Code, Continue, Zed, JetBrains, Amazon Q
• Team sync - organization-scoped projects with role-based access
• Offline mode - local SQLite cache with automatic sync
• Open source - Apache-2.0. Self-host or use the cloud

Next steps

• Quickstart - authenticate and connect in 2 minutes
• Concepts - how memory storage, sync, and team sharing work
• Agent setup guides - configuration for every MCP client
• Agent instructions - make agents reliably use memctl on every turn
• CLI reference - all commands
• MCP tools - full tool and action reference
• Organizations & Teams - roles, projects, and plan limits
• Context System - typed context, bootstrapping, and token budgeting
• Troubleshooting - common issues and fixes