TheGoatPsy/mneme

Overview

A vault-native memory system for Claude Code and MCP clients. It stores memory as markdown files, ensures no LLM is involved in stop decisions, and redacts sensitive data before creating derived stores. Provides accountability by tying memory to a secure vault.

Best for

Best for
Developers using Claude Code who need secure, auditable memory

Use cases

• Persisting long-term context for Claude Code sessions
• Sharing secure, redacted memory across MCP clients
• Auditing memory changes with markdown-based source of truth

Notes

A vault-native memory system for Claude Code and MCP clients. It stores memory as markdown files, ensures no LLM is involved in stop decisions, and redacts sensitive data before creating derived stores. Provides accountability by tying memory to a secure vault.

0 stars on GitHub. Last updated 2026-06-08. Licensed MIT.

Use cases

• Persisting long-term context for Claude Code sessions
• Sharing secure, redacted memory across MCP clients
• Auditing memory changes with markdown-based source of truth

Pros

• Markdown as source of truth keeps memory human-readable and versionable
• Vault integration provides strong accountability and access control
• Automatic redaction protects sensitive information before storage

Cons

• Requires HashiCorp Vault or similar for operation
• Limited to Claude Code and MCP clients today
• No LLM on stop path may reduce flexibility in some workflows

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