Enterprise DNA
M MCP Servers Developer low

AlekseiMarchenko/central-intelligence

by Various

Persistent memory for AI agents. Agents forget. CI remembers.

A

MCP

AlekseiMarchenko/central-intelligence

Added 1 June 2026

Overview

Central-intelligence is a persistent memory system for AI agents, implemented in TypeScript. It allows agents to store and recall information across sessions, addressing the common problem of agents forgetting context.

Best for

Best for
Developers needing a basic, open-source persistent memory layer for AI agents without heavy dependencies

Use cases

  • Maintaining chat history for conversational AI
  • Enabling autonomous agents to persist state across tasks
  • Providing long-term memory for multi-turn interactions

Notes

Central-intelligence is a persistent memory system for AI agents, implemented in TypeScript. It allows agents to store and recall information across sessions, addressing the common problem of agents forgetting context.

1 stars on GitHub. Last updated 2026-05-08. Licensed Apache-2.0.

Use cases

  • Maintaining chat history for conversational AI
  • Enabling autonomous agents to persist state across tasks
  • Providing long-term memory for multi-turn interactions

Pros

  • Directly solves the agent forgetting problem with persistent memory
  • Lightweight TypeScript implementation suitable for many environments
  • Open source with a focused scope

Cons

  • Very early stage with only 1 star, indicating limited community validation
  • May lack advanced features like memory summarization or retrieval-augmented generation
  • Single developer project with potentially sparse documentation

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

Pros

  • Directly solves the agent forgetting problem with persistent memory
  • Lightweight TypeScript implementation suitable for many environments
  • Open source with a focused scope

Cons

  • Very early stage with only 1 star, indicating limited community validation
  • May lack advanced features like memory summarization or retrieval-augmented generation
  • Single developer project with potentially sparse documentation