Echo 06179f94

MCPAI代理

提供方 Pranavj17

AI代理 未标注 无需鉴权 社区 / 第三方 Claude DesktopClaude

ECHO是一个基于AI的角色代理组织模型,通过MCP协议实现自主决策和层级协作,适用于未来科技公司的自动化运营。

一句话选型结论

按资料字段整理

Echo 06179f94 属于AI代理类 MCP 服务器,传输方式未标注,无需鉴权,可直接接入 Claude Desktop、Claude,没有现成安装命令,需按官方文档接入。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类AI代理 传输方式未标注 鉴权无需鉴权 部署方式未标注
GitHub 星标 1 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

AI代理
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

支持客户端

适用场景

Elixir本地部署AI决策组织自动化多代理系统MCP协议本地AI

安装与配置

未标注 无需鉴权
上游未给出传输方式
  1. 1
    准备运行环境

    按官方文档确认运行环境(常见为 Node.js 或 Python),装好后再执行安装命令。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
给代理接外部能力

让代理在任务中途调用搜索、数据库等真实工具。

多步任务编排

把一串操作拆成步骤,由代理按顺序执行。

自动化重复流程

把固定流程交给代理跑,减少手工操作。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 已适配 Claude Desktop、Claude,换客户端不用重新研究配置。
  • 资料标注无需鉴权,装好就能连。

先想清楚的地方

  • 没有现成安装命令,需要按官方文档自己填启动命令和参数。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在AI代理场景里让模型直接调用外部能力的人。
  • 已经在用 Claude Desktop、Claude 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Echo 06179f94当前 未标注 未标注 未标注 无 2026-05-17
Filesystem 560c4363 13 未标注 本地部署 无 2026-05-17
Server Everything (everything) 11 stdio · 本地进程 本地部署 有 2026-05-17
Time 7365cb6d 2 SSE · 服务端推送 远程部署 有 2026-05-18
Everything Claude Code 未标注 stdio · 本地进程 本地部署 有 2026-05-17
Openmemory 557e1585 未标注 stdio · 本地进程 本地部署 有 2026-05-17

服务器档案

AI代理
字段来自项目仓库与公开索引,中文说明由本站标注
提供方Pranavj17上游标注的维护方 Pranavj17
分类AI代理本站按上游分类展示
连接方式未标注上游未给出传输方式
鉴权方式无需鉴权远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · 未标注本地安装时用到的包标识与运行时
部署类型未标注上游给出的部署形态
最近核验2026-05-17上游最后核验时间

详细介绍

7.4 KB
来自项目仓库与上游目录的原文,未做删改

ECHO - Executive Coordination & Hierarchical Organization

An AI-powered organizational model with autonomous role-based agents communicating via MCP protocol

🎯 Vision

ECHO enables future tech companies to operate with AI workers that:
- Make autonomous decisions within their authority
- Collaborate through consensus when needed
- Escalate to appropriate authority levels
- Require human approval for critical decisions
- Communicate naturally across organizational hierarchies

🏗️ Architecture

Each organizational role runs as an independent MCP server that Claude Desktop (or any MCP client) can connect to:```
Claude Desktop / MCP Client
├──> mcp-server-ceo
├──> mcp-server-cto
├──> mcp-server-chro
├──> mcp-server-operations
├──> mcp-server-product-manager
├──> mcp-server-architect
├──> mcp-server-uiux
├──> mcp-server-developer
└──> mcp-server-test-lead

All agents share:
├── PostgreSQL (organizational memory)
└── Redis (message bus)


## 📊 Monitoring Dashboard

**NEW**: Real-time Phoenix LiveView dashboard for monitoring agent activities!```bash
cd monitor
./start.sh
# Open http://localhost:4000

Features:
- 📈 Overview: Daily agent activity summary
- 🔗 Power Delegation: Decision flow tracking
- 📊 Performance: Agent metrics and efficiency
- ⏱️ Timeline: Real-time activity feed

See MONITORING_DASHBOARD_GUIDE.md for details.

🚀 Quick Start

Prerequisites

  • Elixir 1.18+ with Erlang/OTP 27
  • PostgreSQL 16+
  • Redis 7+
  • Ollama (for local AI models)
  • Claude Desktop (for MCP client)

Setup in 3 Steps

1. Start infrastructure:```bash

macOS with Homebrew

brew services start postgresql
brew services start redis

Or use Nix shell (includes all dependencies)

nix-shell


**2. Setup database, LLMs, and agents:**```bash
# Create database and run migrations
cd shared
mix ecto.create
mix ecto.migrate
cd ..

# Install Ollama and download AI models (~48GB)
./setup_llms.sh

# Build all agents
./setup.sh

3. Configure Claude Desktop:

The setup script automatically creates the configuration. Just restart Claude Desktop!

Manual configuration: See CLAUDE_DESKTOP_SETUP.md

Verify Installation```bash

Check system health

./echo.sh summary

Expected output:

● System Status: OPERATIONAL

Infrastructure: ✓ PostgreSQL, ✓ Redis

Agents: ✓ 3 / 9 agents healthy

Test LLM integration for specific agent

./scripts/agents/test_agent_llm.sh ceo

Test all agents' LLM integration

./scripts/agents/test_all_agents_llm.sh


### First Demo

Open Claude Desktop and try:```
Use the CEO agent to approve a strategic initiative:
- Name: "AI Research Lab"
- Budget: $750,000
- Expected outcome: "Advanced AI capabilities"

See DEMO_GUIDE.md for 10 comprehensive demo scenarios.

📚 Documentation

Document Description
CLAUDE_DESKTOP_SETUP.md Connect agents to Claude Desktop
DEMO_GUIDE.md 10 demo scenarios with examples
ECHO_ARCHITECTURE.md Complete system architecture
AGENT_INTEGRATION_GUIDE.md Agent implementation details
LLM_TESTING_SUCCESS.md LLM integration testing guide
OLLAMA_SETUP_COMPLETE.md Ollama and model setup
DISTRIBUTED_SYSTEMS_IMPROVEMENTS.md Reliability & observability
ECHO_SH_README.md Monitoring script documentation

🤖 Available Agents

Each agent has AI assistance via specialized local LLM models:

Agent Role AI Model Status
CEO Strategic leadership, budget allocation qwen2.5:14b ✅ Built + AI
CTO Technology strategy, architecture deepseek-coder:33b ✅ Built + AI
CHRO Human resources, talent management llama3.1:8b ✅ Built + AI
Operations Head Infrastructure and operations mistral:7b ✅ Built + AI
Product Manager Product strategy, prioritization llama3.1:8b ✅ Built + AI
Senior Architect System design, technical specs deepseek-coder:33b ✅ Built + AI
UI/UX Engineer Interface design, user experience llama3.2-vision:11b ✅ Built + AI
Senior Developer Feature implementation, coding deepseek-coder:6.7b ✅ Built + AI
Test Lead Quality assurance, testing codellama:13b ✅ Built + AI

All agents use local AI models via Ollama - zero API costs, complete privacy, works offline.

🎯 Key Features

AI-Powered Decision Making

Every agent has an ai_consult tool for AI-assisted analysis:```elixir

CEO consulting AI for strategic decision

ai_consult(
query_type: "decision_analysis",
question: "Should we expand to European market?",
context: %{
options: ["Immediate expansion", "Pilot program", "Defer"],
budget: "$5M",
timeline: "12 months"
}
)


**Benefits:**
- 🔒 **Private** - All AI runs locally, no cloud APIs
- 💰 **Free** - Zero API costs after setup
- 🎯 **Specialized** - Each role has domain-specific model
- ⚡ **Fast** - Low latency for real-time decisions
- 🌐 **Offline** - Works without internet

### Decision Modes

**Autonomous** - Agent makes decision within authority:```elixir
# CEO can approve budgets under $1M autonomously
approve_budget(amount: 750_000)  # ✅ Autonomous
approve_budget(amount: 5_000_000) # ⬆️ Escalates to human

Collaborative - Multiple agents vote/consensus:```elixir

CTO proposes architecture, team votes

propose_architecture(design: "Microservices")

→ Senior Architect, Operations, Product Manager vote


**Hierarchical** - Escalates up reporting chain:```elixir
# Developer uncertain → Architect → CTO → CEO
escalate_technical_decision(issue: "Database choice")

Human-in-the-Loop - Critical decisions require human:```elixir

Legal, financial, or strategic risks

escalate_to_human(reason: "Regulatory compliance")


### Inter-Agent Communication

Agents communicate via Redis pub/sub + PostgreSQL persistence:```elixir
# Product Manager → CTO
publish_message(
  from: :product_manager,
  to: :cto,
  type: :request,
  subject: "Technical feasibility review"
)

Workflow Engine

Define multi-agent workflows:```elixir
workflow "Feature Development" do
step :product_manager, "define_requirements"
step :senior_architect, "design_system"
step :cto, "approve_architecture"

parallel do
step :senior_developer, "implement_backend"
step :ui_ux_engineer, "design_ui"
end

step :test_lead, "create_test_plan"
step :ceo, "approve_budget"
end


### Health Monitoring

Real-time system observability:```bash
./echo.sh           # Full system status
./echo.sh agents    # Agent health with heartbeats
./echo.sh workflows # Running workflows
./echo.sh messages  # Message queue status
./echo.sh decisions # Pending decisions

🚧 Development Status

Current Phase: Phase 4 - Workflows & Integration

Completed:
- ✅ Phase 1: Foundation (shared library, MCP protocol, database schemas)
- ✅ Phase 2: CEO agent (reference implementation)
- ✅ Phase 3: All 9 agents implemented
- ✅ Phase 4.1: Distributed systems improvements (reliability, observability)

In Progress:
- 🔄 Workflow engine testing
- 🔄 Integration with external systems
- 🔄 Production deployment guides

See ECHO_ARCHITECTURE.md for complete architecture design.

📄 License

MIT License


ECHO - Building the future of AI-powered organizations, one agent at a time. 🚀

常见问题

6 问
答案由本站依据条目字段整理,不引用上游原文
Echo 06179f94 怎么安装?

资料没有给出安装命令,请到官方仓库或文档按说明接入;这类条目多数是生态应用,或尚未发布可安装包。

Echo 06179f94 怎么在 Claude Desktop、Claude 里配置?

它的传输方式未标注,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Echo 06179f94 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Echo 06179f94 需要 API Key 吗?

资料标注为无需认证,装好就能连;如果实际调用时报鉴权错误,按官方文档补上凭据。

Echo 06179f94 支持哪些客户端?

资料标注支持 Claude Desktop、Claude;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Echo 06179f94 能提供哪些能力?

资料显示能力标签包括 AI代理、本地AI、Claude。具体工具清单以官方文档为准。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-17。