Engram Cc
Give Claude Code or Cursor persistent memory across sessions so handoffs, decisions, and codebase context survive restarts.
Overview
Engram-cc is a MCP server for the Build phase that stores, compresses, and retrieves AI coding session memory for handoffs across chats.
What is this MCP server?
- stdio MCP server via npm package @morous-dev/engram-cc (v0.1.4)
- Session handoff so coding assistants pick up where the last chat left off
- SLM compression to shrink noisy history before retrieval
- Semantic retrieval over stored context instead of full transcript replay
- Works as a universal memory layer for AI coding workflows
- Published npm package @morous-dev/engram-cc at version 0.1.4
- stdio MCP transport via official MCP server schema 2025-12-11
- Open-source repo: Morous-Dev/engram-context-continuum on GitHub
Community signal: 1 GitHub stars.
What problem does it solve?
Every new agent session forces you to repeat architecture, todos, and half-finished decisions until context limits erase the thread.
Who is it for?
Indie developers running long-lived Claude Code or Cursor sessions on one repo who need memory outside the default context window.
Skip if: Teams that want hosted team knowledge bases, formal compliance audit trails, or memory with no local MCP setup.
What do I get? / Deliverables
Your coding agent can hand off sessions, pull compressed history, and semantically fetch prior facts so work continues with less re-onboarding.
- MCP tools for session handoff and memory retrieval wired into your agent
- Compressed, searchable history of prior coding-session context
- Reduced re-prompting when resuming work on the same repository
Recommended MCP Servers
Journey fit
Agent memory is listed under Build because it is installed as MCP tooling that augments how you code with agents day to day, even though you benefit in Ship and Operate too. agent-tooling is the canonical shelf for MCP servers that extend the assistant itself rather than a single app feature.
How it compares
MCP memory and retrieval layer, not a markdown note skill or a hosted wiki integration.
Common Questions / FAQ
Who is engram-cc for?
Solo and indie builders using MCP-enabled coding agents who lose context between sessions or hit context limits on active projects.
When should I use engram-cc?
Use it when you repeatedly re-explain the same codebase, want structured handoffs between agent sessions, or need retrieval over past decisions without pasting full logs.
How do I add engram-cc to my agent?
Install @morous-dev/engram-cc from npm, configure the MCP server with stdio transport in Claude Code, Cursor, or another MCP client, then restart the client so tools load.