Nexus RAG MCP Server

Nexus RAG MCP Server

Enables AI agents with long-term memory and retrieval-augmented generation (RAG) capabilities, allowing them to recall past conversations, search local files, and learn user preferences.

Category
访问服务器

README

Nexus RAG MCP Server 🧠

<div align="center">

License Python MCP Test

The "Second Brain" for your AI Assistant.

Configuration • Usage • Tools • Troubleshooting

</div>


🚀 Overview

Nexus is a local Model Context Protocol (MCP) server that gives your AI agents (like Claude Desktop, Cursor, etc.) Long-Term Memory and RAG (Retrieval-Augmented Generation) capabilities.

It runs on your local machine, indexes your files, and allows your AI to:

  1. Recall past conversations and decisions.
  2. Search your local codebase and notes.
  3. Learn your preferences over time.

⚙️ Configuration (Zero-Setup)

You don't need to manually install Nexus. Just use uvx to run it directly from the repository.

1. Claude Desktop App

To use Nexus with Claude, add the following to your config file:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "nexus": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/kaushikkumarkr/RAG-MCP",
        "nexus",
        "serve"
      ]
    }
  }
}

Note: This requires uv to be installed. Nexus will automatically initialize its database at ~/.nexus on the first run.

2. Cursor / Windsurf / Other IDEs

Go to Settings > Features > MCP and add a new server:

  • Name: nexus
  • Type: stdio
  • Command: uvx --from git+https://github.com/kaushikkumarkr/RAG-MCP nexus serve

💬 Usage Examples

Once connected, you can talk to your AI naturally. Nexus will automatically trigger the right tools.

🧠 Memory (Long-Term Storage)

User: "Remember that for the 'Financial Dashboard' project, we depend on the pandas 2.0 library and use pytest for testing."

AI: "Understood. I've saved that context for the Financial Dashboard project." (Tool used: remember)


🔍 Retrieval (RAG)

User: "How do we handle authentication in this codebase?"

AI: "According to docs/auth_flow.md, we use JWT tokens via Auth0..." (Tool used: search_knowledge)


⚡ Context Injection

User: "I'm starting work on the backend API. Get me up to speed."

AI: "Checking project context... Okay, for the backend API:

  1. We use FastAPI.
  2. The DB is Postgres.
  3. Warning: Migration scripts are in /alembic." (Tool used: get_project_context)

🛠️ Available Tools

Tool Description Arguments
search_knowledge Search your local documents. query (str), limit (int)
remember Store a fact or memory. content (str), project (str), tags (list)
recall Retrieve memories. query (str)
get_project_context Get full overview of a project. project (str)
get_user_preferences Get learned user habits. None
ingest_content Save content from other MCPs. content (str), source (str)
add_note Create a markdown note. filename (str), content (str)
list_sources List indexed files. None

🛡️ Citations & Grounding

Nexus citations are precise. Every answer includes:

  1. Source Path: File path (/docs/api.md).
  2. Relevance: Confidence score (0-1).
  3. Quote: Exact text chunk used.

❓ Troubleshooting

Server not starting?

Run the built-in status check to verify your database and config:

nexus status

"Tool not found" error?

Ensure you have restarted your client (Claude/Cursor) after editing the config file.

How do I see what's happening?

Nexus logs to stderr. In Claude Desktop, you can open the Developer Console to see the MCP logs in real-time.


🏗️ System Architecture

Nexus uses a Dual-Layer Cognitive Architecture that separates fast retrieval from deep understanding.

System Design Architecture

Built with: Python 3.11 • Qdrant • Sentence-Transformers • RAGAS

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选