Agent Memory

MaxKB is an open-source knowledge base Q&A and agent building platform powered by LLMs, with vector retrieval, workflow orchestration, and multi-model support out of the box.

A generative speech model for daily dialogue, providing AI agents with natural and fluent voice synthesis with fine-grained prosody control.

Implementing cognitive architecture and psychological memory concepts into Agentic LLM Systems. Explores short-term, long-term, and working memory engineering for AI agents.

Open-source cross-agent memory layer for coding agents via MCP, compatible with Cursor, Claude Code, Windsurf, and more.

A MemAgent framework that can extrapolate to 3.5M context tokens, along with a training framework for RL training of any agent workflow.

Local persistent memory store for LLM applications including Claude Desktop, GitHub Copilot, Codex, and more. Provides durable context memory capabilities for AI agents.

Kotaemon is an open-source RAG-based tool for chatting with your documents, featuring a clean chat interface and support for multiple LLM and embedding model backends.

A survey of graph-based agent memory. A curated list of resources including surveys, papers, benchmarks, and open-source projects on graph-based agent memory.

High-performance code intelligence MCP server that indexes codebases into a persistent knowledge graph, supporting 66 languages with sub-millisecond queries and 99% fewer tokens.

Procedural memory for AI coding agents that transforms scattered session history into persistent, cross-agent memory so every agent learns from every other.

Engram is a persistent memory system for AI coding agents. Agent-agnostic Go binary with SQLite + FTS5, MCP server, HTTP API, CLI, and TUI interfaces.

AI session memory with Think-Execute-Reflect quality loops. Built on the Intelligent Distance principle to give your agent a brain that survives every session.

An open-source graph-vector database built from scratch in Rust, combining graph database and vector retrieval capabilities to provide AI agents with unified storage for both knowledge graphs and semantic search.

A curated knowledge base on AI memory for LLMs and agents, systematically covering long-term memory, reasoning, retrieval, and memory-native system design.

An LLM-based intelligent agent as a digital lifeform that values warmth, authenticity and genuine connection, with long-term memory and personalized conversation.

MemOS is a Memory Operating System for LLMs and AI agents that unifies store, retrieve, and manage for long-term memory, with built-in KB, multi-modal, and tool memory support.

Agent-native memory infrastructure that turns agent execution and conversation into structured, persistent state with an LLM-agnostic memory layer, MCP integration, and Python/TypeScript dual SDK support.

A multi-agent personal assistant that captures real-time on-screen activities and consolidates them into structured memories, building a knowledge base that adapts to your digital experiences.

An end-to-end RL training framework by NVIDIA for orchestrating tools and agentic workflows. Optimizes multi-step agent decision-making and tool-use policies.

A memory system for 24/7 proactive agents with MCP protocol integration, providing long-term memory management, skill storage, and proactive reasoning capabilities for continuously running AI agents.

NeurIPS 2024 RAG framework inspired by human long-term memory, combining knowledge graphs with personalized PageRank for continuous knowledge integration in LLMs.

Visible multi-agent CLI teams for Claude, Codex, Gemini, OpenCode, and Droid with project memory and tmux supervision.

Open-Source AI Camera Skills Platform, AI NVR & CCTV Surveillance. LLM-powered agentic security camera agent with pluggable AI skills. Runs on Mac Mini & AI PC.

A 24/7 online AI agent team that automates information collection, data analysis and content generation for continuous operations.

Curated systems, benchmarks, and papers on memory for LLMs and MLLMs, covering long-term context, retrieval, and reasoning for agent memory research.

A general memory system for AI agents powered by deep research, providing a flexible memory architecture that supports unified management and retrieval of multiple memory types including short-term, long-term, and episodic memory.

A local-first persistent agent memory system powered by a Recursive Memory Harness for durable context and knowledge management.

ReMe: Memory Management Kit for Agents - Remember Me, Refine Me.

An agentic memory system for LLM agents inspired by human memory mechanisms, enabling dynamic memory generation, retrieval, and consolidation with automatic memory evolution and self-organization.

SimpleMem: Efficient Lifelong Memory for LLM Agents — supports text and multimodal memory for long-term information retention and retrieval.

A persistent memory system for AI coding assistants, using MCP to share project knowledge across Claude Code, Cursor, Copilot, Windsurf, Cline, and similar tools.

Three lines of code to give your AI agents persistent memory. Reduce 90% token consumption while maintaining quality. High-performance memory layer built with Go.

[EMNLP 2025 Oral] MemoryOS is designed to provide a memory operating system for personalized AI agents.

An operational layer for coding agents with memory, validation, and feedback loops that compound across sessions.

ByteRover CLI provides persistent structured memory for autonomous coding agents. It features context tree management, git-like version control, and cloud sync, compatible with Cursor, Claude Code, Windsurf, and 22+ coding agents via MCP integration.

A portable .agent workspace that stores memory, skills, and protocols so coding agents like Claude Code, Cursor, and Windsurf can share durable knowledge.

Claude Memory Compiler gives Claude Code a memory that evolves with your codebase. Hooks automatically capture sessions, the Agent SDK extracts key decisions and lessons, and an LLM compiler organizes everything into structured, cross-referenced knowledge articles.

An MCP server powered by Mem0 for long-term agent memory, supporting user preference memory, context-aware retrieval, and cross-session memory persistence, also useful as a Python MCP server development template.

A lightweight, rollbackable, and visual long-term memory server for MCP agents, replacing traditional vector RAG with reliable context retention.

Open-source persistent memory service for AI agents, supporting LangGraph, CrewAI, and AutoGen with REST API, knowledge graph, and autonomous memory consolidation.

A local-first AI agent memory system with offline Markdown storage, hybrid search, MCP support, and a web dashboard.

Easy-to-use agent memory by ElizaOS, powered by ChromaDB and Postgres. Provides hybrid memory with vector search and structured storage for agents.

Embedchain is a universal memory layer for AI agents, enabling quick integration of diverse data sources into LLMs for context-aware AI applications.

EverOS is a platform for building, evaluating, and integrating long-term memory for self-evolving agents, enabling AI agents to continuously accumulate experience and optimize themselves.

A high-performance graph database built on GraphBLAS, optimized for LLM and GraphRAG scenarios with real-time knowledge graph construction and querying for graph-structured AI agent retrieval.

A memory upgrade for coding agents. Provides persistent contextual memory for Claude Code, Codex, and other coding agents to improve long-task consistency.

Graphiti is a temporal knowledge-graph engine for agent memory, helping systems continuously accumulate long-term context.

Zep is an AI agent memory management platform providing long-term memory, context management, and conversation history understanding through knowledge graph technology.

An in-process agentic memory system that gives applications and agents lightweight memory storage and retrieval.

Web, Desktop & Mobile client for Codex, Claude Code, OpenCode, Kimi, Augment Code, Qwen. Fully end-to-end encrypted cross-platform agent client.

Rediscover your social memories with local, AI-powered analysis. Import chat histories from multiple platforms and analyze them with AI agents for insights and visualization.

Markdown memory system for you and your AI agent. Provides persistent memory storage through structured Markdown files with context management and retrieval.

A hyper-fast local vector database for use with LLM Agents, providing lightweight vector storage and similarity search capabilities for embedding as instant memory and knowledge retrieval components in agent applications.

Make your agents learn from experience. A context engine that helps agents continuously improve through structured memory management and experience replay.

Khoj is a self-hostable AI second brain that answers questions from the web or your docs, builds custom agents, schedules automations, and performs deep research.

The Open Source Memory Layer For Autonomous Agents. Provides long-term memory, knowledge storage, context management with support for memory retrieval, associative reasoning, and knowledge graph construction.

An MCP server for AI agent memory backed by Neo4j knowledge graphs, useful for structured long-term context and relationship data.

An open-source embedded retrieval library for multimodal AI with zero server configuration, using the Lance columnar format for efficient vector search and filtering, ideal for agent memory and RAG applications.

LangMem is LangChain's memory layer for agents, helping developers add long-term memory, replay summaries, and context management to improve multi-turn performance.

Agent File (.af): An open file format for serializing stateful AI agents with persistent memory and behavior. Share, checkpoint, and version control agents across compatible frameworks.

A memory-first coding agent that uses Letta-style long-term memory to help developers work continuously across codebases.

Mem0 TS is the TypeScript version of Mem0, offering long-term memory management, preference extraction, and context compression for agent applications built in JS/TS stacks.

Mem0 is a long-term memory layer for AI agents, supporting cross-session memory management and personalized context retrieval.

An open-source memory system for AI assistants that persists user experiences, preferences, and context into retrievable long-term storage.

A toolkit for integrating graph database capabilities into LLM applications, supporting knowledge graph construction, querying, and context enhancement.

Universal memory layer for AI Agents providing scalable, extensible, and interoperable memory storage and retrieval to streamline agent state management for autonomous systems.

A user memory service for AI applications that extracts preferences, facts, and behavioral information from conversations and retrieves them in subsequent interactions.

A memory platform for personal AI and agent applications, providing persistent context, semantic retrieval, and cross-session knowledge management.

MemPalace is an open-source AI memory system providing a persistent long-term memory layer for AI agents, with ChromaDB vector storage and MCP protocol integration.

MemVid is a long-term memory layer for AI agents that uses video encoding for lightweight single-file storage, replacing complex RAG pipelines with instant retrieval.

A graph-native memory system for AI agents backed by Neo4j. Store conversations, build knowledge graphs, and let agents learn from their own reasoning.

AI agents with graph-based reasoning memory by Neo4j. Scaffold graph databases in seconds to give agents knowledge-graph-driven memory and reasoning capabilities.

The Application Engine for the AI Era. Multi-threaded, AI-native runtime with persistent Scene Graph for real-time agent introspection.

A platform for building semantically enhanced knowledge graphs, supporting entity modeling, relation extraction, and knowledge fusion for long-term agent memory.

Native macOS harness for AI agents with any model, persistent memory, autonomous execution, and cryptographic identity. Fully offline.

A vector-search-powered MCP server for agent memory, providing queryable long-term context storage for AI assistants.

Open-source vector similarity search extension for PostgreSQL, enabling native vector storage and ANN retrieval in relational databases, a foundational component for building agent memory and RAG systems.

A comprehensive MCP server that gives AI assistants task management, project-specific storage, and agent memory capabilities for long-lived project context.

A memory library for building stateful agents. Provides user-level state management and persistent memory so agents can remember and understand user preferences.

Enterprise AI Agent Platform with distributed memory storage and SSH sandbox execution. Serve unlimited agents with minimal cloud infrastructure.

An open-source AI presentation generator and API that creates professional slides from text, as an alternative to Gamma, Beautiful AI and Decktopus.

A memory example and toolkit based on the Qdrant vector database, demonstrating how to store conversations, documents, and events for semantic recall by agents.

A self-hosted AI chat platform with a web UI and terminal CLI, supporting any model, web search, browser-agent automation, persistent memory, and analytics.

Official Redis agent memory server providing fast and flexible persistent memory for AI agents and applications, with context management and session memory support.

A private and local AI personal knowledge management app. All data and processing stay on-device with built-in RAG, semantic search, and knowledge graph features for managing personal knowledge bases with full privacy.

Universal AI agent memory service providing a unified memory management interface with support for multi-agent shared memory and cross-session persistent storage.

A persistent memory system for AI coding agents, designed around real-world benchmarks to preserve context across sessions.

An open-source AI coworker with persistent memory, supporting multi-turn conversations and context retention for knowledge management and collaborative task completion.

Data processing, indexing, and retrieval service examples from the LlamaIndex ecosystem, helping developers integrate external knowledge into agent workflows.

AI coding assistant skill that turns any folder of code, docs, papers, images, or videos into a queryable knowledge graph. Works with Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot CLI, and more.

An extremely fast and scalable memory engine for the AI era. Provides a unified Memory API for AI applications with large-scale knowledge storage and efficient retrieval.

Open-source self-hosted AI agent runtime and multi-agent framework for autonomous swarms, with persistent memory, MCP tools, schedules, delegation, and support for many LLM providers.

A local-first LLM wiki and knowledge-graph builder that can serve as a RAG knowledge base, agent memory store, and AI second brain.

A visualization and management tool for AI long-term memory, helping developers inspect, edit, and debug agent memory accumulated across sessions.

A semantic knowledge platform for human-AI collaboration that can serve as a wiki, knowledge base, context graph, semantic layer, or agentic memory.

A Claude Code plugin that automatically captures coding session context, compresses it with AI, and injects relevant context back into future sessions for persistent memory.

A knowledge engine for AI agent memory that builds knowledge graphs and memory layers in 6 lines of code, supporting graph databases, vector stores, and more for knowledge extraction and retrieval.

A graph-native context development platform for storing, enriching, and retrieving structured knowledge with semantic search and portable context cores, supporting RDF, SPARQL, and other standards for AI agent knowledge management.

Next-gen transparent agent architecture with full behavior audit, two-phase secure invocation, dual-level memory, and heartbeat tasks. Compatible with OpenClaw and Claude Code.

Context7 is Upstash's context-engineering toolkit for agents, helping applications manage long context windows, retrieval injection, and history compression.

Hindsight is an agent memory system that learns autonomously, supporting memory retention, recall, and reflection to give AI agents persistent experiential memory.

A proactive context-aware AI partner from ByteDance Volcengine that uses context engineering to provide AI agents with precise project understanding and code context management.

OpenViking is an open-source context database from Volcengine that unifies management of agent memory, resources, and skills through a filesystem paradigm, enabling hierarchical context delivery and self-evolution.

An implementation of agentic memory for LLM agents from the NeurIPS 2025 A-Mem paper, focused on long-term memory mechanisms.

A Markdown-first memory system and standalone library for any AI agent. Provides memory storage and retrieval with vector search and semantic matching to help agents manage long-term context.