# NLWeb

> Open protocols and tools for conversational website interfaces via MCP

NLWeb is an open source reference implementation by Microsoft for adding natural language querying to websites using Schema.org metadata. It natively acts as a Model Context Protocol (MCP) server, allowing both human users and AI agents to query structured site data through standardized natural language endpoints.

Website: https://github.com/microsoft/NLWeb

## What it is for

- Add a conversational search interface to a content or e-commerce website
- Expose website content directly to AI agents via Model Context Protocol (MCP)
- Convert natural language queries to SQL against enterprise data stores using Schema.org ontologies
- Route user queries to appropriate LLMs based on cost and quality thresholds

## Features

- Native Model Context Protocol (MCP) server integration using standard 'ask' endpoints
- JSON responses formatted using standardized Schema.org vocabularies
- Pre-built AskAgent query agent with sample web UI and data ingestion tools
- Multi-vector database support including Qdrant, Milvus, Postgres, and Azure AI Search
- LLM provider support for OpenAI, DeepSeek, Gemini, Anthropic, and Hugging Face
- DataFinder module for natural-language-to-SQL translation across enterprise schemas
- ModelRouter and NLWebScorer for dynamic model selection and search result ranking

## Pricing

Open source

## Details

- Type: GitHub repo
- Category: [Development](https://xiry.app/topics/development.md)
- Open source: yes
- Platforms: Linux, macOS, Windows, Self-hosted, API
- For: Web developers and AI engineers building agent-ready websites and conversational search interfaces
- Tags: mcp, python, schema org, ai agents, search, conversational ai

## Why it was saved

- [@omarsar0, 23 May 2025](https://x.com/omarsar0/status/1925900575666733207): Lets developers turn any website into an AI app using standardized natural language APIs; NLWeb natively speaks MCP, so the same APIs serve both humans and agents. Positions itself as a foundational protocol layer for the AI web ('MCP is to NLWeb what HTTP is to HTML'), leveraging schema.org and RSS as the semantic layer.

## Related

- [Composio](https://xiry.app/resources/composio): Tooling and authentication layer connecting AI agents to external applications (GitHub repo, Freemium)
- [RAGLite](https://xiry.app/resources/raglite): Lightweight Python toolkit for RAG using DuckDB or PostgreSQL (GitHub repo, Open source)
- [Scrapling](https://xiry.app/resources/scrapling): Adaptive Python web scraping framework with anti-bot bypass and element relocation. (GitHub repo, Open source)
- [Firecrawl](https://xiry.app/resources/firecrawl-1984): Turn websites and documents into clean, LLM-ready markdown or structured data. (Website, Freemium)
- [awesome-llm-apps](https://xiry.app/resources/awesome-llm-apps): Open-source collection of runnable AI agent, skill, and RAG templates (GitHub repo, Open source)
- [Parse](https://xiry.app/resources/parse): Turn any website into a structured REST API using natural language (Web app, Freemium)

Updated 10 October 2026. Page: https://xiry.app/resources/nlweb

Make your own library: Xiry takes your saved reels, tweets, Reddit bookmarks and Watch Later videos, makes them searchable and pulls the useful stuff out, like this library. It's invite-only for now; join the waitlist at https://xiry.app/#waitlist.