I built an MCP Server to stop AI from hallucinating Quranic verses – Introducing quran-search-engine-mcp

​I wanted to share a tool I’ve been working on to solve a major issue when using LLMs for religious study: Hallucinations.

We’ve all seen AIs confidently misquote verses or mix up chapters. To fix this, I developed quran-search-engine-mcp. It’s a Model Context Protocol (MCP) server that connects any MCP-compatible AI (like Claude Desktop) directly to a reliable Quranic database.

​How it works:
​Instead of letting the AI guess or "recall" a verse from its training data, the AI sends the search intent to the MCP server. The server fetches the literal, accurate text, and the AI then uses that verified data for its response.

​Features:
​Zero Hallucinations: The AI only interprets the search results; the text itself is pulled directly from the source.
​Semantic/Contextual Search: It doesn't just look for exact words. For example, if you search for "Prophet Yunus and all his synonyms," the engine understands the context and retrieves verses mentioning "Yunus," "Dhul-Nun," and "The Companion of the Whale."
​Plug & Play: If you use Claude Desktop, you can add it instantly.

​Quick Setup for Claude Desktop:
​Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "quran-search-engine-mcp": {
      "command": "npx",
      "args": ["-y", "quran-search-engine-mcp"]
    }
  }
}

Why I built this:​This opens up possibilities for building highly accurate educational tools, research assistants, or any project that requires verified Quranic data without the risk of AI-generated errors.
​I’d love to get your feedback on the implementation!

www.github.com/adelpro/quran-search-engine-mcp