linklore/linklore-mcp

Overview

MCP server that gives AI agents persistent project memory. Agents record decisions, pitfalls, and specs during work, then recall them in later sessions with code context. Written in Python and built for agent-to-agent use rather than human interaction.

Best for

Best for
Developers building agentic coding workflows that need continuity across sessions.

Use cases

• Record architectural decisions as they are made
• Retrieve past pitfalls when working in the same codebase
• Store and recall project specs across agent sessions

Notes

MCP server that gives AI agents persistent project memory. Agents record decisions, pitfalls, and specs during work, then recall them in later sessions with code context. Written in Python and built for agent-to-agent use rather than human interaction.

1 stars on GitHub. Last updated 2026-08-06.

Use cases

• Record architectural decisions as they are made
• Retrieve past pitfalls when working in the same codebase
• Store and recall project specs across agent sessions

Pros

• Persistent memory across sessions
• Context-aware recall tied to code
• Open source and language-agnostic via MCP

Cons

• Very early stage (1 GitHub star)
• No human-facing interface by design
• Requires MCP-compatible agent setup

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