Mcp Memento
Give Claude Code, Cursor, or other MCP hosts persistent memory so project context survives across separate coding sessions.
Overview
Memento is an MCP server for the Build phase that provides persistent memory and cross-session context management for AI coding agents.
What is this MCP server?
- PyPI package mcp-memento (v0.2.31) with stdio MCP transport per server schema 2025-09-16
- Persistent memory and cross-session context management for long indie builds
- Open-source server at github.com/annibale-x/mcp-memento
- Host-agnostic: any MCP client that supports stdio registration
- Cuts repeated re-briefing on stack choices, conventions, and in-flight tasks
- Server version 0.2.31 on PyPI identifier mcp-memento
- MCP server schema https://static.modelcontextprotocol.io/schemas/2025-09-16/server.schema.json
- Single published package entry with stdio transport type
Community signal: 2 GitHub stars.
What problem does it solve?
Every new agent session starts cold, so solo builders keep re-uploading the same project facts, preferences, and open threads.
Who is it for?
Multi-day agent-assisted builds on one codebase where you want lightweight persistent memory without building a custom RAG pipeline.
Skip if: One-shot codegen with no MCP setup, or teams that already enforce all context in external PM/docs tools and will not connect an MCP memory server.
What do I get? / Deliverables
After you register mcp-memento, your MCP host can read and write durable memory so conversations and implementation work continue with shared context instead of from scratch.
- Registered stdio MCP server io.github.annibale-x/mcp-memento in your agent host
- Persistent memory layer agents can use across multiple coding sessions
- Reduced manual re-briefing on stable project context and ongoing tasks
Recommended MCP Servers
Journey fit
Persistent agent memory is infrastructure you wire in while building and maintaining a product with an AI coding assistant, not a one-off launch or research task. It extends the agent stack (stdio MCP on PyPI) rather than shipping features, SEO, or ops dashboards—canonical home is agent-tooling under Build.
How it compares
Stdio MCP memory server, not a markdown Claude skill pack or a curated skills marketplace entry.
Common Questions / FAQ
Who is mcp-memento for?
Solo and indie builders using MCP-enabled agents (Claude Code, Cursor, Claude Desktop, Windsurf-style hosts) who need memory that survives across sessions on the same project.
When should I use mcp-memento?
Use it when you are actively building or iterating with an agent and find yourself repeating project context, conventions, or task state every time you open a new chat.
How do I add mcp-memento to my agent?
Install the PyPI package mcp-memento (0.2.31), add a stdio MCP server entry pointing at that package in your client’s MCP configuration, restart the host, and confirm the server exposes memory tools per the repo README.