Memory Mcp
Give Claude Code episodic and entity memory across sessions so your agent remembers project decisions, people, and tasks without re-pasting context every time.
Overview
io.github.chenxiaofie/memory-mcp is a MCP server for the Build phase that provides episodic and entity persistent memory for Claude Code over stdio.
What is this MCP server?
- Episodic plus entity memory model for long-running Claude Code projects
- PyPI package chenxiaofie-memory-mcp v0.2.1 with stdio transport
- Optional CLAUDE_PROJECT_ROOT env scopes memory to a repo path
- Targets 为 Claude Code 提供持久化记忆—persistent memory as first-class MCP
- stdio local server—memory stays tied to your project machine layout
- Version 0.2.1 on PyPI registry
- stdio transport; one documented optional env CLAUDE_PROJECT_ROOT
- Dual memory model: episodic (情景) + entity (实体) per description
Community signal: 11 GitHub stars.
What problem does it solve?
Long build cycles reset agent context, so you repeat architecture explanations and lose track of entities the agent already learned.
Who is it for?
Solo developers living in Claude Code who want local stdio memory MCP without building a custom vector store.
Skip if: Teams needing enterprise audit trails, multi-user shared memory with RBAC, or memory without Python MCP hosting.
What do I get? / Deliverables
Claude Code can read and write durable episodic and entity memory scoped to your project root, shrinking onboarding cost each session.
- Running stdio memory MCP wired into Claude Code
- Reusable episodic and entity memory records across agent sessions
Recommended MCP Servers
Journey fit
Build is when context volume explodes; persistent memory MCP reduces rework as you integrate features and agents over weeks. Agent-tooling is the shelf for MCP servers that extend the coding agent itself—memory fits alongside other Claude Code infrastructure.
How it compares
Project-scoped agent memory MCP, not a general documentation search or RAG pipeline skill.
Common Questions / FAQ
Who is io.github.chenxiaofie/memory-mcp for?
Claude Code users who need episodic and entity memory persisted across sessions on a single project.
When should I use io.github.chenxiaofie/memory-mcp?
Use it during build (and ongoing operate work) when context reuse matters—multi-week agents, recurring entities, and repeated stack explanations.
How do I add io.github.chenxiaofie/memory-mcp to my agent?
Install chenxiaofie-memory-mcp from PyPI, configure stdio in Claude Code MCP, optionally set CLAUDE_PROJECT_ROOT to your repo, then restart Claude Code.