Remembr – Persistent memory infrastructure for AI agents

Summary: Remembr is an open-source memory server that provides AI agents with persistent, searchable memory across sessions using Redis for short-term caching and PostgreSQL with pgvector for long-term episodic storage. It includes SDKs for Python and TypeScript and native adapters for eight AI frameworks.

What it does

Remembr enables AI agents to retain memory between sessions by combining Redis session cache with PostgreSQL-based long-term storage and supports hybrid semantic search. It offers multi-tenant isolation and GDPR-compliant data deletion.

Who it's for

Developers building AI agents who require persistent memory and integration with popular AI frameworks like LangChain, AutoGen, and LlamaIndex.

Why it matters

It solves the problem of AI agents forgetting information between sessions by providing a structured memory architecture beyond simple vector search indexes.