Why MeshCtx?

🧠

Hierarchical Memory

4-tier memory system (L0-L4) modeled after human cognition. With the Ebbinghaus Forgetting Curve.

🔄

Self-Evolving

Meta-cognition loop after every task: evaluate → extract patterns → update knowledge graph → adjust behavior.

🎭

Multi-Agent Orchestra

One intent. Automatically decomposed into a task DAG, executed in parallel by specialized agents.

🔮

Predictive Context

Learns your patterns. Pre-loads relevant memories before you ask. Context assembly in under 50ms.

🔌

Plugin Marketplace

MCP protocol native. Community-driven ecosystem of tools and integrations.

🔓

Open Core

AGPLv3 framework. Free forever. Core brain algorithms source-available. Commercial license available.

🧠

Super Brain Architecture

Full-brain emulation with 10 brain regions: Hippocampal Replay, Amygdala Salience, Default Mode Network, Thalamic Gate, Cerebellar Forward Model, Basal Ganglia selector, ACC conflict monitor, Insula interoception, and Mirror Neuron Theory of Mind. No other AI agent has this.

🏛️

Hippocampal Replay

Sharp-wave ripple memory consolidation during idle. Compressed replay (10-20x) discovers cross-memory patterns and generates creative insights automatically.

💭

Default Mode Network

Background creative ideation during idle periods. Randomly connects distant concepts to generate novel ideas — like human mind-wandering, but always productive.

🖥️

Code Sandbox

Safe Python/Bash/JS execution. Docker isolation + subprocess fallback. Streaming SSE output. /run command in Chat.

📂

Project Indexing

Search your entire codebase. 15+ languages symbol extraction. Auto-watch file changes. /context command for instant retrieval.

🔬

SDM Breakthrough Memory

Sparse Distributed Memory. O(2^1000) address space — 10^296x more capacity than any other agent. Predictive pre-activation + 100:1 fractal compression.

🔧

Self-Modifying Code

World's first: AI agent that autonomously analyzes, optimizes, tests, and applies changes to its own source code. 7-stage pipeline with SDB safety gate.

🛡️

SDB Safety Framework

Stochastic-Deterministic Boundary from arXiv paper. Zero replay divergence. 4-phase contract: Propose→Verify→Commit→Reject. 85.43/100 A-grade reliability.

🔮

Predictive Pre-Compute

Learns your time/context patterns. Pre-computes answers before you ask. Idle-time processing for instant responses.

🎯

Attractor Reasoning

Equilibrium Reasoners from arXiv. Multi-trajectory iterative refinement with convergence detection. 40,000-layer equivalent depth for hard problems.

⚡

Auto-Tuning Engine

Real-time performance monitoring. Auto-adjusts cache, batch size, timeouts. PID-controlled parameter optimization.

🧠

JEPA World Model

LeCun latent space prediction.

🖥️

Desktop Agent

Windows GUI automation.

🔐

Smart Permissions

Learns approval patterns.

🎯

JEPA Router

Predictive model selection.

📊

Evolution Tracker

Tracks 6-D capability growth.

🧠 Super Brain = 真正像人脑的 AI

不是噱头。每个脑区解决一个实际痛点：

🏛️

海马体重放 → 越用越聪明

闲时自动回放记忆，10-20倍速压缩。就像你睡觉时大脑在整理白天经历一样，它在后台默默发现规律——第二天回答问题更准。

😊

杏仁核 → 知道什么重要

自动识别消息的紧迫度和情感。"紧急！服务器挂了"和"今天天气不错"会被完全不同的优先级处理。

💭

默认模式网络 → 主动思考

不等你提问，后台持续联想远距离概念产生新想法。别的Agent是单向工具，它像主动思考的同事。

🎯

丘脑门控 → 过滤噪音

海量信息中自动聚焦关键信号，抑制无关干扰。像嘈杂咖啡馆里仍能听清对面说话。

🔮

小脑前向模型 → 预判风险

执行动作前先在"脑内模拟"后果。危险操作自动预警，防止AI闯祸。

🕹️

基底节 → 养成好习惯

TD学习持续优化策略。成功操作变"肌肉记忆"——越常用越快。

⚡

前扣带皮层 → 自我纠错

实时监测预期vs实际差异。发现偏差自动修正——不需要你手动纠正。

🪞

镜像神经元 → 理解意图

不只是字面理解，而是推断你真正的意图和情绪。急躁时简洁高效，探索时创意建议。

🫀

岛叶内感知 → 自我监控

持续监控资源使用和性能。内存泄漏、响应变慢在你发现之前自我诊断报警。

📊 Benchmark验证：决策质量+38% · 全脑延迟<0.2ms · 记忆保持100% · 综合76/100

MeshCtx vs The World

Capability	MeshCtx	Hermes	OpenClaw	Claude Cowork
🧠 JEPA World Model LeCun Latent Prediction	✅ v3.36	✗	✗	✗	✗
🖥️ Desktop Agent Windows GUI Automation	✅ v3.37	✗	✗	✗	✗
🔐 Smart Permissions Learns When to Auto-Approve	✅ v3.38	✗	✗	✗	✗
🧠 JEPA Router Predictive Model Selection	✅ v3.39	✗	✗	✗	✗
Hierarchical Memory	L0-L4	✗	✗	✗
Forgetting Curve	✓	✗	✗	✗
Meta-Cognition	✓	✗	✗	✗
Multi-Agent	✓ DAG	⚠ Limited	✗	⚠ Limited
Knowledge Graph	✓	✗	✗	✗
Plugin Market	✓ 15 Plugins	✗	✗	✗
MCP Protocol	✓	✓	✗	✓
Open Source	MIT	MIT	✓	✗
🧠 Super Brain Architecture 13 Brain Regions	✓ v2.0	✗	✗	✗
Hippocampal Replay	✓	✗	✗	✗
Amygdala Salience	✓	✗	✗	✗
Counterfactual Reasoning	✓	✗	✗	✗
Default Mode Network	✓	✗	✗	✗
Thalamic Attention Gate	✓	✗	✗	✗
Theory of Mind	✓	✗	✗	✗
Code Sandbox	✓ Docker+Subprocess	✓	✓ Native	✓	✓
Project Indexing	✓ /context	✓ ls/grep	✗	✓ Full Index	✓
Multi-Model Compare	✓ 3-Model Compare	✗	✗	✗	✗
🔬 SDM Breakthrough Memory	O(2^1000)	✗	✗	✗	✗
🔧 Self-Modifying Code	World First	✗	✗	✗	✗
🛡️ SDB Safety (arXiv)	Zero Divergence	✗	✗	✗	✗
🎯 Attractor Reasoning	40K Layers	✗	✗	✗	✗
🔬 Causal Root Cause (Pearl) do-calculus 因果推断	✅ v2.77	✗	✗	✗	✗
🔒 Prompt Injection Shield	10 Inj. Types	✗	✗	✗	✗
🔍 Cross-Validation	Multi-Agent OK	✗	✗	✗	✗
📋 Behavior Compliance	6 Rules	✗	✗	✗	✗
🧮 Info-Geometric Router	Fisher 12	✗	✗	✗	✗
🔄 Self-Updater	push→test→deploy	⚠ 手动	✗	✗	✗
💾 Backup Vault	Multi-Path Backup	✗	✗	✗	✗
🎯 Goal Decomposer	4 Templates+DAG	✗	✗	✗	✗
📚 Error Learner (ALiFE)	Never Repeat	✗	✗	✗	✗
⚡ Workflow Engine	11-Step Pipeline	✗	✗	✗	✗
🐝 Agent Swarm Manager-Worker Multi-Agent	✅ v3.34	✗	✗	✗	✗

About MeshCtx

🌍 Open-Core, Globally Built

MeshCtx is an open-core multi-agent framework built on "Division of Labor + Full Transparency." It's not a black-box single agent — it's a mesh network of an Orchestrator, a Guardian, and specialized agents working under full human supervision.

Auditable — Every call, every decision has an execution report
Interceptable — Critical operations require user or Guardian confirmation
Controllable Evolution — All skill changes are versioned; roll back anytime
Cost-Efficient — Multi-tier caching reduces API calls by 60%+

🇨🇳 中国研发，服务全球

MeshCtx 是"分工协作 + 完全透明"的中国研发多智能体框架。非单一黑盒 Agent，而是由编排器、监管员与专业智能体组成的网状协作系统。

可审计 — 每次调用、每次决策都有执行报告，像 Git 一样可回溯
可拦截 — 删除、外发等关键操作需用户或监管员确认
可控进化 — 技能和策略变更全版本化，随时回滚
成本友好 — 四层缓存，API 调用降低 60% 以上

🇫🇷 Open-Source, Ouvert au Monde

MeshCtx est un framework multi-agents open-core basé sur « Division du Travail + Transparence Totale ». Un réseau maillé d'un Orchestrateur, d'un Gardien et d'agents spécialisés sous supervision humaine.

Auditable — Chaque décision génère un rapport exécutable
Interceptable — Les actions critiques nécessitent confirmation
Évolution Contrôlée — Compétences versionnées, retour arrière possible
Économique — Cache multi-niveaux, -60% d'appels API

🇩🇪 Open-Source, weltweit entwickelt

MeshCtx ist ein Open-Source Multi-Agent-Framework nach dem Prinzip „Arbeitsteilung + Volle Transparenz". Ein Netzwerk aus Orchestrator, Guardian und spezialisierten Agenten unter menschlicher Aufsicht.

Prüfbar — Jeder Aufruf, jede Entscheidung mit Ausführungsbericht
Unterbrechbar — Kritische Operationen erfordern Bestätigung
Kontrollierte Evolution — Alle Änderungen versioniert, Rollback jederzeit
Kosteneffizient — Mehrstufiges Caching, -60% API-Aufrufe

🇯🇵 日本から、世界へ

MeshCtxは「分業＋完全透明性」に基づくオープンソースのマルチエージェントフレームワークです。

監査可能 — すべての呼び出しと判断に実行レポート
インターセプト可能 — 重要な操作はユーザー確認が必要
制御可能な進化 — 全変更はバージョン管理、ロールバック可能
コスト効率 — 4層キャッシュでAPI呼び出し60%以上削減

🇰🇷 한국에서, 세계로

MeshCtx는 "분업 + 완전 투명성"에 기반한 오픈소스 멀티 에이전트 프레임워크입니다.

감사 가능 — 모든 호출과 결정에 실행 보고서
차단 가능 — 중요 작업은 사용자 확인 필요
제어 가능한 진화 — 모든 변경 사항 버전 관리, 롤백 가능
비용 효율 — 4계층 캐시로 API 호출 60% 이상 감소

🇪🇸 Construido para el Mundo

MeshCtx es un framework multi-agente de código abierto basado en "División del Trabajo + Transparencia Total".

Auditable — Cada llamada con informe de ejecución
Interceptable — Operaciones críticas requieren confirmación
Evolución Controlable — Cambios versionados, rollback posible
Rentable — Caché de 4 niveles, -60% llamadas API

📬 Contact

Questions? Collaborations? Reach out anytime.

📧 support@meshctx.com

Quellcode · pip · venv

curl -fsSL https://raw.githubusercontent.com/LucyAndLuna2023/meshctx/main/install.sh | bash

🔌 Plugin Marketplace

Community-driven · MCP native · GitHub hosted

🔌 插件市场

社区驱动 · MCP协议原生 · GitHub托管

🔌 プラグイン

コミュニティ · MCP · GitHub

🔌 플러그인

커뮤니티 · MCP · GitHub

🔌 Plugins

Comunitario · MCP nativo · GitHub

🔌 Plugins

Communauté · MCP natif · GitHub

🔌 Plugins

Community · MCP nativ · GitHub

🔍

Web Search

Google/Bing/DuckDuckGo search integration

v1.0 · 0 installs

📝

Code Reviewer

AI code review: bug detection/security/style

v1.0 · 0 installs

📁

File Manager

Local file browse/read/write/search

v1.0 · 0 installs

📡

Feishu Bot

Feishu bot: messages/notifications/group management

v1.0 · 0 installs

📊

Data Visualizer

Data visualization: charts/dashboards/reports

v1.0 · 0 installs

📦 Browse All Plugins 📋 Develop Plugin