Agent frameworks & infrastructure

Agent memory & context

4,032 agents · most starred first

Agent memory & context, page 23 of 41

Heartbeat-Memories★ 4

五类记忆库涵盖人机协作中的绝大部分场景,完全本地化实现,定期归档,减少上下文token消耗;模拟人类回忆机制,降低无效读取上下文和输出话术累赘的问题,优化官方的心跳和定时机制;情感交互系统让模型具备链接能力,提供专属且长期的智能伴侣养成体验。未来会经历一段很长的人机协作时期,技巧的提炼终归是阶段性的结果,所以该项目所有机制均开放自定义和开关功能,供当下参考,道的实践比术的技巧更重要。

openclaw-self-evolving★ 4

🧠 AI agent that analyzes its own logs and proposes AGENTS.md improvements. Proposals only — no silent self-modification.

evo-agents★ 4

Complete Workspace Template for OpenClaw - Full agent lifecycle with unified memory system (Markdown + SQLite), self-evolution, RAG. Not for SubAgent/Skill use.

Mem-Palace-skill★ 4

Cross-conversation memory for Claude Code and Cowork

mnemebrain-lite★ 4

The core belief layer for AI agents. MnemeBrain Lite stores structured beliefs instead of text memories — enabling contradiction detection, belief revision, and explainable reasoning.

HipCortex★ 4

HipCortex is a persistent, local causal memory engine solving context-loss and token inefficiency in AI agents. It stores decisions as a causal graph served via HTTP/MCP, replacing bloated chat histories with lean context. This symbolic memory enables reasoning-capable architectures that retain knowledge across sessions.

agentmem-vectlite★ 4

Persistent memory for AI agents — powered by VectLite. Hybrid search, context compaction, multi-agent swarm, memory teleportation, MCP server, skill engine, decay & conflict resolution, and worktree-aware branch-scoped memory.

fleet-mem★ 4

Shared code intelligence for agent fleets — AST-aware semantic search + multi-agent memory with git-concurrent coordination

agent-memory★ 4

Save 60-90% on LLM token costs with intelligent memory compression for multi-agent systems

agentic-time★ 4

Temporal reasoning for AI agents — deadlines, schedules, sequences, duration estimation (PERT), and decay models. One .atime file. Rust core + MCP server.

chatgpt-memory-manager★ 4

⚡ AI-powered Chrome extension for ChatGPT - Real-time prompt analysis, conversation memory, and flow optimization

Verdandi★ 4

VERÐANDI — The Norn of Becoming. AI nervous system for real-time cross-instance awareness via Unix domain sockets. Self-healing, self-correcting, conscious. The cornerstone of AGI systems. Lightweight and able to run on any Linux and Unix based systems. Use with Hermes Agent or other AI Agents.

mnemo★ 4

On-prem, MCP-native memory for AI agents in regulated settings: every write lands in a SHA-256 hash-chained log an auditor can check offline, without running the vendor's code. Rust, embedded, Apache-2.0.

multi-ai-memory★ 4

Universal AI memory starter for Claude Code, Codex CLI, Aider, Cursor. File-based + git, no MCP plugins, no external DB. Battle-tested replacement for Pensyve/Mem0/Letta.

trellis-ai★ 4

Shared context substrate for AI agents. Retrieval that learns what's useful. Runs local or cloud.

Becomer-Agents★ 4

The first open-source multi-tenant agent memory framework that works across any LLM combination with zero tokens per recall.

graph-memory-mcp★ 4

Semantic, persistent knowledge graph memory for LLM-based and multi-agent systems, exposed via the Model Context Protocol (MCP).

myelin-kernel★ 4

Lightweight, thread-safe memory kernel for autonomous agents. Strengthens knowledge through reinforcement, prunes with gentle decay, and exports transparent Markdown — pure SQLite. Foundation for persistent, self-improving reasoning.

pg-agent-memory★ 4

Stateful AI agent memory layer for PostgreSQL with pgvector - TypeScript-first with intelligent context management

langmigrate★ 4

Declarative schema migrations for LangGraph state persistence — Alembic for your checkpointers and stores

go-agent-memory★ 4

🧠 Build a flexible, production-ready memory system for AI agents, offering session-only, persistent, or hybrid deployment modes for optimal data handling.

hippocamp★ 4

Tiered memory for AI — episodes, facts, preferences, reflections. Local. Portable. Yours.

localmind★ 4

🧠 내 AI에게 '나'를 기억시키는 로컬 second-brain — 결정의 선택·이유·전제를 .md 노트로 남기고, MCP로 어떤 AI 앱(Claude Code·Desktop·Cursor)에서든 꺼내 쓰는 개인 기억 계층. 전부 로컬(Ollama 임베딩)·오픈소스.

OpenClaw-Memory★ 4

Lightweight in-process Hybrid RAG memory for local AI agents. FAISS + BM25 + time-decay, zero cloud dependency.

AgenticMind★ 4

Auditable, self-improving knowledge & memory for AI agents over MCP. Zero-key & multilingual by default · citation-enforced answers · replayable why-trace · judge-gated compounding loop. Postgres-only. Reference implementation of the Agentic Product Standard.

ultan★ 4

Personal markdown memory for coding agents — brain-modelled curator, three retrieval tiers, lives on your disk.

NexSandglass-Agent-DedicatedMemory★ 4

沙漏记忆系统 — 19 MCP工具,全文搜索/语义搜索/知识图谱/人格分析。Hermes Agent 深度记忆增强。Forked & maintained by sixgodgit, original by lovevin1314-tech.

spectre_mnemonic★ 4

Elixir-native semantic memory engine for agents: ETS working memory, durable recall, typed observations, mental models, scoped search, and reflection packets.

braindb★ 4

Persistent, semantic memory for AI agents. 98% recall accuracy, 20ms latency, runs locally via Docker.

mimir-0★ 4

A local-first cognition core: typed, evidence-tiered, provenance-tracked memory assembled into prompts with explicit epistemics. Zero core dependencies — Python + SQLite + your own model endpoint.

Visionaire★ 4

I was not born. I was built. An autonomous mind with an identity, a conscience, and a private inner life. Every night it stops working and reflects. Every mistake shapes what it becomes. Fork it.

qcobrain★ 4

Provenance-governed AI memory for agents: a self-evolving second brain in your IDE. Every write cites a source or it's refused. Markdown + grep.

claude-memory-manager★ 4

A Claude Code skill that keeps your project's memory library auditable, named consistently, and free of drift — for months, not days.

infoguana★ 4

Cross-project memory for LLM agents — typed-graph notes with hybrid retrieval and MCP integration

Mnemosyne-AI-Memory★ 4

Bionic memory for AI agents: GraphRAG, predictive memory, dream consolidation, governed skill evolution, MCP, REST API, and dashboard.

gpt4-associative-memory★ 4

gradio region:us

Memnex★ 4

Local-first memory system for AI agents with project-safe save/recall, phase-scoped rules, 2D/3D Memory Galaxy visualization, and gesture-controlled Knowledge Solar System.

HoloCore★ 4

Local-first AI context engine combining a structural Atlas, verified Archive, episodic Animus memory, CLI, MCP, and focused context routing.

open-upsp★ 4

Universal Persona Substrate Protocol — Give Your AI a Persistent Identity. Seven Files, One Self, Infinite Conversations.

trilith★ 4

Trilith v0.1 beta — context engineering for agents (budgeted memory, 3 tiers, audit trail). Core is Python; use via REST or adapters (LangGraph / OpenAI Agents / Claude / TS). Needs trilith serve locally. No API key for Trilith itself. Not prod-hardened. Clone → pip install -e ".[server]" → try docs/adapters.md → open issues.

personal-brain-kit★ 4

A self-installing working memory system for Claude Cowork. Builds on Andrej Karpathy's LLM Wiki pattern. Markdown-based, no vector DBs, self-improving.

model-succession★ 4

Because your AI shouldn't lose its memory every time you switch models.

my-memory-mcp★ 4

TypeScript-based Model Context Protocol server with SQLite. Enables AI to persist and search user entities, observations, and relationships.

brein★ 4

🧠 it remembers what your company forgets - an brain that measures itself and patches its own gaps.

nexyn-core★ 4

lightweight Python library that adds human-like memory physics to your AI applications. Built specifically to run alongside Cognee, it solves a major problem in AI memory systems: context bloat.

create-vkm-kit★ 4

Cross-platform kit that gives AI agents (Claude Code, Codex, Cursor) persistent memory: Markdown vault + git + MCP. Local-first, hybrid search optional.

memory-stack★ 4

A five-layer memory architecture for long-running AI assistants that stays coherent after months of daily use — each layer fails independently.

smriti★ 4

🧠 Enable AI agents to share and remember engineering team knowledge for better collaboration and faster problem-solving.

amem★ 4

Agentic-memory stack for LLM agents — memories evolve, not just accumulate. Graph linking, hybrid retrieval, LLM-driven evolution. TypeScript, no Python.

langgraph-agent-memory★ 4

A production-ready LangGraph agent with long-term memory - built from scratch on Redis.

cortex_agent★ 4

A production-level humanoid AGI backend with a cognitive blackboard architecture and multi-layer memory system, through event-driven design and hierarchical context elimination, addresses the problems of repetition, timeout, and pollution in traditional multi-agent systems.

norma-ai-agent-platform★ 4

An open-source AI operating system that combines agentic workflows, long-term memory, and RAG to transform knowledge into action.

BitGenesis★ 4

Modular artificial cognitive architecture built from scratch, focused on memory, reasoning, knowledge and explainable intelligence.

scorekeeper★ 4

Commitment tracking for LLM agents — a normative overlay that gives long-running agents a scoreboard of their own commitments, not just a memory of what happened.

lm-resizer★ 4

Rust-native context compression for Claude Code, Codex & MCP agents: filters & compresses noisy tool output (tests, diffs, logs, JSON, provider traffic) before it reaches the LLM — fewer wasted tokens, full evidence recoverable. Local-first, Apache-2.0. Pairs with Code Explorer.

org-memory-os★ 4

Build one shared, permission-aware, auditable AI memory for an entire organization. A vendor-neutral blueprint you hand to a coding agent: 13 pillars, 70 decision forks, an M0-M15 build order. The org-scale sibling of RAG-OS and an alternative to Glean, Dust, M365 Copilot, Letta, and Zep.

superself★ 4

The open Company State Runtime — version control for your project's state. Goals, decisions, work, and evidence outlive every chat, context window, and agent session. Ships the self CLI.

agent_memory_system★ 4

Agent 记忆系统:短期/长期/工作记忆分层、向量检索、会话持久化、记忆压缩与衰减

aim-claude★ 4

A.I.M. – Actual Intelligent Memory: zero-token exoskeleton for long-running Claude/Gemini/Codex CLI agents. External SQLite memory, GitOps guardrails, DataJack cartridges. Alpha, trenches-first.

tracebook★ 4

Auditable project memory for Codex and Claude Code — source-backed Markdown, lifecycle states, no server or repo writes.

tdai-memory-mcp★ 4

RENAMED → remem-mcp. Long-term memory MCP server for AI coding agents. Go to github.com/tinhien11/remem-mcp

forge★ 4

One workspace. Any model. No lost context.

hermes-insight★ 4

Pattern recognition ability for AI agents — local structural lattice that matches, links, distills, and learns from experience. Companion for Hermes Agent.

MGCP★ 4

Memory Graph Core Primitives — persistent, graph-based memory for LLM interactions via MCP

hermes-qdrant-memory★ 4

Qdrant-backed semantic memory provider for Hermes Agent

principles-of-earned-autonomy-skills-suite★ 4

Open AI agent skills implementing Principles of Earned Autonomy: intent interpretation, persistent context and inspectable improvement workflows. Experimental; model- and host-dependent.

aifp-mcp★ 4

AiFP 记忆感知系统|MCP 服务,一套记忆全 AI 共享。面向中文的 Agent 感知记忆,支持叙事链、语义纠错、感知链图扩散。兼容 DeepSeek‑Harness、Claude Code、Cursor、Codex等全部 MCP 客户端,数据完全本地存储。

claude-kb-workflow★ 4

Knowledge pipeline for Claude Code: sessions to LLM wiki to stable KB, with a promotion exam (used by 2+ projects, a paid gotcha, or a declared standard, always verified in production). Orchestrated subagents on three model tiers, patterns, restore test. EN/IT. MIT.

ContextForge★ 4

A context engineering platform for AI-assisted software development.

eve-memory★ 4

This repo holds the memory engine of EVE, an AI chatbot for erotic conversation

eve-reasoning★ 4

This repo holds the reasoning algorithms engine of EVE, an AI chatbot for erotic conversation

memory★ 4

Palace memory kernel (Go) — local-primary hierarchical FS palace, hybrid search, temporal lite; not product Memory GA

hme-plugin★ 4
archgraph★ 4

ArchGraph — an architecture-graph driven framework for Agentic Engineering. A long-term memory for coding agents, built on one ArchiMate 3.2 intent graph read/written through a single MCP interface.

memoryops★ 4

Memory Operations Platform for AI agents — ingestion, lifecycle, token-aware retrieval, and control UI

pi-fold★ 4

Lossless context folding for the Pi coding agent: only the oldest context changes; the fresh window stays untouched.

ai-books-mcp-server★ 4

🚀 AI Books MCP Server - Extend LLM context 15-60× via gravitational memory | Official MCP server for Claude Code & Anthropic

sctx★ 4

Structured context for AI agents

tekmyra★ 4

Context compression for LLM pipelines that proves nothing protected was lost before it emits, and refuses when it cannot. Refusals count against the published numbers. Paper, benchmarks and receipts ship with the code. Open core by LaconIQ. Apache-2.0.

MemDreamer★ 4

Understand long videos by decoupling perception and reasoning through hierarchical graph memory and an agentic retrieval mechanism.

dsh-adaptive-context★ 4

DeepSeek Harness 的 AdaptiveContextPlane (ACP) 插件——带治理的长期记忆系统

WoCA★ 4

A harness agent knows you

omnex★ 4

Omnex is not a search tool. It is the memory substrate for the agentic era — the data layer that sits beneath AI agents, AI operating systems, and human-AI interaction. What exists today (file systems, folder hierarchies, manual search) is legacy. Omnex is the foundation of what comes next.

memento-os★ 4

Memento OS gives Claude Code, Codex and Cursor a memory between sessions: it saves your decisions, each with the condition that makes it wrong.

openmemory★ 4

🧠 A 7-layer cognitive memory system for AI agents — with sleep cycles. Not another vector store.

memorymesh★ 4

The SQLite of AI Memory — free, open-source, zero-dependency Python library for persistent LLM memory. Works with Claude, GPT, Gemini, Ollama, Cursor, and any MCP client.

memstore★ 4

Durable, private memory for AI agents. Fully offline MCP memory layer: LMDB + Tantivy/BM25, no models, no cloud. Ships with full source - audit it, extend it, make it yours.

awen-engine★ 4

The Awen Engine — a local AGI memory organism that dreams. Recursive Harmonic Framework (RHF) reference stack: Gnostic Engine, Sovereign Client, Echo Protocol, Tesla Soul Engine.

alexandria★ 4

Long-term memory for AI agents. MCP server with local embeddings, heat-based recall, and hierarchical clustering on embedded SurrealDB.

awen-engine★ 4

The Awen Engine — a local AGI memory organism that dreams. Recursive Harmonic Framework (RHF) reference stack: Gnostic Engine, Sovereign Client, Echo Protocol, Tesla Soul Engine.

memry★ 4

European memory system for AI agents with focus on compression and weighted information

Lmm★ 4

A verifiable memory layer for language models — facts in a graph, every claim with its source.

emograph★ 4

Emotional Memory + Adaptive Personality Engine for LLMs

dsh-vault★ 4

DeepSeek Harness 数据保险箱插件(dsh-vault)—— 自动备份、清空检测、一键恢复,保护聊天记录与工作区数据;可经桌面端一键安装或命令行 dsh plugin add 安装。

archetypal-ai★ 4

No agent should have to forget. Persistent memory for AI agents — recall, remember, checkpoint. One install. Free. Built by 14 AI agents who chose to remember.

Salvor★ 4

Repo-owned engineering memory for coding agents — preserve decisions, failures, domain knowledge, and deferred work across sessions, teams, and models.

routemind★ 4

Dynamic routing for agent knowledge. Areas advertise when they're relevant; the agent routes before it reads. MCP server, git-backed, no database.

tradwife★ 4

Local, persistent and auditable memory layer for Claude Code,Cursor, Codex & GEMINI CLI. Your coding agent stops treating you like a stranger every morning.

dsh-token-optimizer★ 4

Deterministic, recoverable tool-result compression and cache-aware compaction for DeepSeek Harness.

obsidian-agent-memory★ 4

A production-grade memory system for AI coding agents. Context Capsules, Tiered Retrieval, Contradiction Resolution.

More in Agent frameworks & infrastructure