MCPFast / Tools / Self-hosted AI memory server with MCP and evidence provenance

GitHubMCP★★★★☆

Self-hosted AI memory server with MCP and evidence provenance

A self-hosted AI memory server using MCP, featuring evidence provenance, conflict detection, embeddings, and PostgreSQL/Neo4j support.

View on GitHub

Self-Hosted AI Memory Server with MCP and Evidence Provenance

For AI developers building complex, stateful agents, managing memory and ensuring data integrity is paramount. This self-hosted AI memory server, available on GitHub, offers a robust solution for storing, retrieving, and verifying AI-generated information. It leverages the Message Passing Convention (MCP) for efficient communication and incorporates advanced features like evidence provenance and conflict detection to maintain a reliable knowledge base.

What it Does

This tool functions as a dedicated memory server for AI agents. It allows for the storage of factual statements, observations, and reasoning chains generated by your AI models. Crucially, it tracks the origin and context of each piece of information through evidence provenance, enabling you to understand how specific memories were formed. This is essential for debugging, auditing, and building trust in AI systems. The server also includes mechanisms for detecting and resolving conflicting information, ensuring a consistent and accurate memory state.

Key Features

Who it's For

This tool is designed for AI developers and researchers who are building sophisticated AI agents requiring persistent, verifiable memory. It is particularly useful for projects involving: