mnemoverse/mcp-memory-server

Overview

A hosted memory server for AI agents that learns from feedback and fades recall by recency. It provides a shared memory layer accessible across Claude, Cursor, VS Code, and ChatGPT via a single API key. Listed on the official MCP Registry.

Best for

Best for
Developers building multi-tool AI workflows that need persistent, feedback-aware memory

Use cases

• Persist conversation context across different AI tools for continuity
• Rerank memory importance based on user feedback
• Manage agent recall in multi-session workflows

How to use

Tools exposed

• memory_write
• memory_read
• memory_feedback
• memory_stats
• memory_delete
• memory_delete_domain

Tested with

Claude Code, Cursor, VS Code, Windsurf, Zed, JetBrains, Cline, Continue

Example client config

[object Object]

Notes

A hosted memory server for AI agents that learns from feedback and fades recall by recency. It provides a shared memory layer accessible across Claude, Cursor, VS Code, and ChatGPT via a single API key. Listed on the official MCP Registry.

1 stars on GitHub. Last updated 2026-06-18. Licensed MIT.

Use cases

• Persist conversation context across different AI tools for continuity
• Rerank memory importance based on user feedback
• Manage agent recall in multi-session workflows

Pros

• Cross-platform memory sharing with one key reduces setup overhead
• Feedback-driven reranking improves relevance over static storage
• Official MCP listing suggests community vetting

Cons

• Early-stage project with only 1 star on GitHub
• Dependence on hosted service which may have uptime or latency
• Unclear documentation for use beyond basic integration

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.