Natural Language Web
NLWeb, short for Natural Language Web, aims to be the fastest and easiest way to effectively turn your website into an AI app. A natural language interface for websites using the model of their choice and their own data. Every NLWeb instance is also a Model Context Protocol (MCP) server, allowing websites to make their content discoverable and accessible to agents and other participants in the MCP ecosystem if they choose.
How does it work?
NLWeb leverages semi-structured formats like Schema.org, RSS and other data that websites already publish, combining them with LLM-powered tools to create natural language interfaces usable by both humans and AI agents. The NLWeb system enhances this structured data by incorporating external knowledge from the underlying LLMs for richer user experiences.
How do I get started?
The NLWeb GitHub repo contains everything you need to get started:
- The lightweight code that controls the core service to handle natural language queries, as well as documentation on how this can be extended and customized.
- Connectors to some of the most popular models and vector databases, as well as documentation to add other models of your choice.
- Tools for adding your data in Schema.org, JSONL, RSS and other formats to your chosen vector database.
- A web server frontend for the service and a simple UI that allows users to send queries to the web server.
source: https://news.microsoft.com/source/features/company-news/introducing-nlweb-bringing-conversational-interfaces-directly-to-the-web/
Written by Microsoft Corporate Blogs Published May 19, 2025
Model Context Protocol
The Model Context Protocol, or MCP for short, is a standard for connecting AI assistants to the systems where data resides.
MCP lets AI models draw data from sources like business tools and software to complete tasks, as well as from content repositories and app development environments.
MCP enables developers to build two-way connections between data sources and AI-powered applications (e.g., chatbots). Developers can expose data through “MCP servers” and build “MCP clients” — for instance, apps and workflows — that connect to those servers on command.
source: https://techcrunch.com/2024/11/25/anthropic-proposes-a-way-to-connect-data-to-ai-chatbots/
Transforming the Web with Natural Language: My NLWeb Presentation at Nashua CLOUD .NET & DevBoston
View the slides on SlideShare: https://www.slideshare.net/slideshow/transform-any-website-into-a-conversational-experience-with-nlweb/281034902
What Is NLWeb?
NLWeb (Natural Language Web) is a robust protocol and toolset developed by Microsoft that turns any traditional website into a conversational interface, leveraging the power of large language models. It’s built around the Model Context Protocol (MCP), allowing developers to process natural-language queries and respond using structured Schema.org JSON.
In my session, I demonstrated how NLWeb works, highlighting its design for flexibility (enabling the swapping out of models, vector databases, and embeddings), and how it seamlessly connects to data and APIs to deliver intelligent, real-time responses to users.
Real-World Impact
I also highlighted real-world use cases where NLWeb is already in action:
- Tripadvisor – enabling users to plan family trips through conversation
- Eventbrite – allowing event discovery through natural-language search
- O’Reilly, Qdrant, Delish, Shopify, and others – showcasing early success in turning structured content into AI-driven UX
These examples demonstrate how businesses are already leveraging the potential of conversational web interfaces to drive engagement and discovery.
How to Get Started
For those interested in experimenting or building with NLWeb, here are a few resources I shared:
- GitHub: https://github.com/microsoft/NLWeb
- Quick start guide:
docs/nlweb-hello-world.md - Local test interface:
http://localhost:8000/static/debug.html - Azure deployment:
docs/setup-azure.md
Whether you’re a developer, architect, or product leader, NLWeb offers a modern and modular approach to embedding LLM-driven intelligence into any web property.
No comments:
Post a Comment