In this video, I look at Web MCP and how it could have a huge impact on how agents operate on the web and work with websites.

Site: https://www.openresponses.org/
HF: https://huggingface.co/blog/open-responses
Demo:  https://github.com/samwit/Open-Responses-Demo

For more tutorials on using LLMs and building agents, check out my Patreon
Patreon: https://www.patreon.com/SamWitteveen
Twitter: https://x.com/Sam_Witteveen 

🕵️ Interested in building LLM Agents? Fill out the form below
Building LLM Agents Form: https://drp.li/dIMes

👨‍💻Github:
https://github.com/samwit/llm-tutorials

⏱️Time Stamps:
00:00 Intro
00:28 Google Chrome: WebMCP Early Preview
01:14 Origin of MCP
05:07 Agent-User Site Interaction: Three Pillars
06:31 Chrome 146 Release
08:05 Structure Interactions for the Agentic Web

Sam Witteveen AI is a publication offering insights, tutorials, and resources for artificial intelligence (AI) enthusiasts and practitioners. Readers can learn about machine learning algorithms, deep learning frameworks, and AI applications. With tutorials, case studies, and expert interviews, Sam Witteveen AI provides  guidance and expertise for building and deploying AI solutions.

Sam Witteveen

Google Chrome has shipped an early preview of WebMCP, a new standard that allows websites to expose structured tools directly to AI agents. Instead of agents scraping HTML or processing screenshots (which consumes thousands of tokens), they can now call functions on web pages. The standard includes two APIs: declarative (for standard HTML forms) and imperative (for complex JavaScript interactions). This enables agents to interact with websites more efficiently through structured tool calls rather than simulating user interactions, potentially reducing token costs significantly. The feature is currently behind a flag in Chrome and expected to roll out more broadly in coming months.

The Rise of WebMCP