Real-time-web-search-RAG-with

A conversational agent that combines static knowledge (RAG over your docs) with real-time web search using an MCP server.

Installation

Run this command in your terminal to add the MCP server to Claude Code.

Run in terminal:

View docs

Command

claude mcp add --transport stdio pranithchowdary-real-time-web-search-rag-with-mcp python -m src.mcp_server.server \
  --env OPENAI_API_KEY="your-openai-api-key" \
  --env SERPAPI_API_KEY="your-serpapi-key (if using SerpAPI for web search)" \
  --env CUSTOM_VECTORMAP="path or identifier for local vector store, if applicable" \
  --env AZURE_OPENAI_API_KEY="optional, if using Azure OpenAI"

How to use

This MCP server powers the real-time hybrid retrieval system that combines local document search (RAG) with real-time web search when needed. Start the MCP server to enable external tool or web API calls as fallback sources for answers that cannot be fully resolved from your local knowledge base. The server coordinates between the local vector store (e.g., FAISS/Chroma) and external search tools, returning a unified response with citations from both sources. To use it, run the MCP server and then query via your preferred interface (REST API or your Python client); if the local data is insufficient, the MCP layer will automatically invoke the web search tool or scraper to fetch fresh information and enrich the answer.

How to install

Prerequisites:

Python 3.8+
Git
Optional: Docker (for containerized deployment)

Clone the repository git clone https://github.com/PranithChowdary/Real-time-web-search-RAG-with-MCP.git cd Real-time-web-search-RAG-with-MCP
Create a virtual environment and activate it python -m venv venv // On Windows use: venv\Scripts\activate source venv/bin/activate
Install dependencies pip install -r requirements.txt
Configure environment variables
- Copy the example env file: cp .env.example .env
- Add your API keys and any required settings (OPENAI_API_KEY, SERPAPI_API_KEY, etc.)
Run the MCP server (as described in the mcp_config section or using the CLI) python -m src.mcp_server.server
(Optional) Run Docker (if preferred) docker-compose up -d

Notes:

See docs/DEPLOYMENT.md for production deployment and security considerations.
The sample environment file (.env.example) shows typical keys to configure.

Additional notes

Tips and notes:

Ensure your API keys are kept secure and not committed to VCS.
If you rely on vector stores, periodically refresh or update your local indexes to improve accuracy.
The MCP integration will surface sources/citations for each answer; verify external sources when precision is critical.
If web search results are noisy, you can adjust the orchestration logic in your local code to favor RAG results and only fallback to MCP for certain query patterns.
For debugging, check logs for MCP calls and ensure that the web/search tools are reachable from your execution environment.

Related MCP Servers

chunkhound

1.1k

Local first codebase intelligence

VectorCode

811

A code repository indexing tool to supercharge your LLM experience.

mcp-pinecone

148

Model Context Protocol server to allow for reading and writing from Pinecone. Rudimentary RAG

langgraph-ai

LangGraph AI Repository

Archive-Agent

Find your files with natural language and ask questions.

mcp-raganything

API/MCP wrapper for RagAnything