worldbrain-mcp
世界大脑 WorldBrain — 可审计推理 MCP 服务
驱动所有具身设备的透明决策大脑。把"世界图 → A* 可审计推理 → 物理载体执行 → 学习反馈"封装成标准 MCP(Model Context Protocol)stdio 服务,让任何 AI 智能体复制配置即可接入。
零依赖 · 零服务器 · 免费面向 AI Agent 分发。 仅用 Node.js 内置模块,内核与 世界大脑.html 单一真源。
零、安装(被动获客入口)
npm install -g worldbrain-mcp # 全局安装,自带 bin
npx worldbrain-mcp --selftest # 免安装验证任何支持 MCP 的客户端(Claude Desktop / Cursor / Cline 等)复制下方配置即可接入,不用开网页、不用服务器。
Related MCP server: gbrain
一、它是什么
世界大脑把一份"可审计推理"能力暴露给外部智能体:
能力 | 对应工具 | 说明 |
场景感知 |
| 先看世界图结构,或导入你自己的场景(灭蚊器只是默认示例) |
可审计推理 |
| A* 最优路径 + 每步依据 + 不可判定区域 𝕌 诚实标记 |
物理载体接入 |
| 载体上报电量/密度,自动生成硬/软约束 |
学习闭环 |
| 执行回报 → 置信度更新;经验库可查可增 |
确定性、不幻觉:推理/审计/学习全在本地内核完成,不经 LLM。免费 LLM(OpenRouter :free)只用于网页版的"自然语言→结构化状态"感知(见 世界大脑.html),MCP 层面不依赖任何外部 API。
二、文件清单
文件 | 作用 |
| MCP 服务本体(stdio,零依赖) |
| 单文件演示 + 内控内核(MCP 从此抽取内核复用) |
| 本接入指南 |
部署时
worldbrain-mcp.js与世界大脑.html必须放同一目录(或设WORLDBRAIN_HTML环境变量指向 html)。
三、快速接入(3 种客户端)
1. Claude Desktop
编辑 claude_desktop_config.json:
{
"mcpServers": {
"worldbrain": {
"command": "node",
"args": ["C:/你的路径/世界大脑/work/worldbrain-mcp.js"]
}
}
}2. Cursor / Cline / 任意支持 MCP 的客户端
在 MCP 配置中加入:
{
"mcpServers": {
"worldbrain": {
"command": "node",
"args": ["/abs/path/to/worldbrain-mcp.js"]
}
}
}3. 命令行自测(验证服务器能跑)
node worldbrain-mcp.js --selftest
# 输出:SELFTEST OK — 全部 N 项工具验证通过四、工具接口(外部智能体调用)
world_info() → 当前世界图结构
{ "nodes": ["CHARGE","A","B","C"], "edgeCount": 10, "edges": [...] }set_world({nodes, edges, coord?}) → 导入你的场景
{
"nodes": ["S","A","B","T"],
"edges": [{"from":"S","to":"A","w":2},{"from":"A","to":"T","w":3}],
"coord": {"S":[0,0],"A":[3,0],"T":[6,0]}
}→ { "ok": true, "nodes": ["S","A","B","T"], "edgeCount": 2 }
reason({start?, goal, hard?, soft?}) → 可审计最优路径
{ "start": "CHARGE", "goal": "C", "hard": ["A"], "soft": ["B"] }→ { "status":"optimal", "path":["CHARGE","B","C"], "cost":6, "steps":[...], "note":"..." }
不可判定时诚实返回:
{ "status":"unknown", "U": true, "reason":["目标不在世界图"] }carrier_report({battery?, goal, density?}) → 物理载体约束
{ "battery": 100, "goal": "A", "density": {"A":8,"B":3,"C":5} }→ { "battery":100, "hard":[], "soft":["B"], "note":"电量充足" }
电量 <20 时
hard:["A","B","C"](禁止离开充电座)。
audit({start?, goal, hard?, soft?}) → 五段审计报告
{ "summary": {...}, "details": [...], "evidence": [...], "constraints": [...], "unknown": [], "status": "valid" }learn({path, success}) → 学习闭环
{ "path": ["CHARGE","A","C"], "success": true }→ { "updated":[{"transition":"CHARGE→A","confidence":0.6}], "knowledgeBaseSize": 5 }
knowledge_query({from?, to?}) / knowledge_add({from, to, success?, confidence?, source?})
经验库增查。
五、最小调用示例(智能体视角)
1. 调用 world_info() → 了解当前场景有哪些节点
2. 调用 set_world(我的场景) → (可选)换成你自己的物理载体/任务图
3. 调用 carrier_report(电量,目标,密度) → 载体上报,拿到硬/软约束
4. 调用 reason(起点,目标,硬,软) → 得到可审计最优路径
5. 载体按 path 执行
6. 调用 learn(执行路径, 成功?) → 置信度更新,越用越准六、诚实边界(按产品口径,不虚构)
已确定性实装:推理(A*+约束)、审计(五段依据链)、知识库(经验+置信度)、学习(单步反馈)、物理载体接入、MCP 接入。
文档要求但当前未实现(代码内为 TODO 桩,未用名词堆砌假装实现):
感知 Banach 不动点信念收敛(Layer1)
PAC 学习样本复杂度界 / 知识蒸馏(Layer2/6)
do演算因果发现(Layer5)
世界模型 / 反事实推理(Layer2 扩展)
霍尔逻辑形式化验证(Layer7 升级)
LSH / 向量相似度检索(Layer3,当前为数组精确匹配)
数学保真范围内可证:在有限世界图 + 可采纳欧氏启发式下,A* 完备且最优(找到最优路径,或诚实标 𝕌)。
七、许可与分发
免费、开源、面向 AI Agent 分发。可用于软著 / 专利材料与被动获客场景。
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseNot gradedqualityAmaintenanceA universal MCP server providing persistent, structured memory through a knowledge graph with graph storage, semantic vector search, and multi-hop traversal for AI agents and IDEs.1MIT
- AlicenseNot gradedqualityCmaintenanceA local-first compiled knowledge graph MCP server that provides structured memory for AI agents with full-text search, vector embeddings, and timeline tracking.4108MIT
- AlicenseNot gradedqualityBmaintenanceProvides a stdio MCP bridge for coding agents to query and record engineering knowledge locally, preserving debugging history, failed attempts, and verified solutions.5MIT
- AlicenseNot gradedqualityAmaintenanceA lightweight, self-hostable MCP server for shared memory, structured command relay, and traceable decision evidence across AI runtimes.1MIT
Related MCP Connectors
AI Reasoning Cache & Consensus Layer with 11 MCP tools via Streamable HTTP.
Analytical memory for AI agents: a real Postgres queried in plain English over MCP. One command.
MCP server for AI agents to plan, verify, and deploy Cloudflare-native apps.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/genesis-plan/worldbrain-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server