In these very early days of what just might be the equivalent of a new industrial revolution, my current thinking can be summed up a little like this:
AI should enable human productivity. It is not a replacement for human creativity.
Just as a single example, here’s a guide I wrote very recently so that you can prototype interesting ideas for content research, market research, fact-checking and data journalism.
So much is possible, especially for people with ideas who perhaps previously couldn’t execute without immediate support from “technical” people. On this note, I think AI levels the playing field and gives us all a chance to achieve better things, faster at work.
My Background
I’m Richard Baxter, and I help technically-minded marketers navigate the shift toward an agent-driven web. As a Marketing Technologist, my work revolves around the protocols that govern how AI interacts with our brands, specifically MCP, UDP, and the rising Universal Commerce Protocol. Houtini is where I demystify these technical surfaces, offering practical workflows and insights to help both developers and marketers leverage the productivity potential of an AI-native workplace.
I founded a digital agency called Builtvisible way back in 2009. With help from some very talented people, I grew it into a globally recognised agency specialising in technical SEO and content strategy. Eventually, Builtvisible was acquired by Brave Bison in 2025.
My career started in usability studies at a credit card bank, moved into e-commerce site launches, and eventually led to founding Builtvisible. Throughout, the constant has been a focus on technical infrastructure – whether that was teaching site architecture, keyword research, MS Excel and duplicate content in the early SEO days, or now MCP server integration and local LLM deployment.
Since then, my day-to-day work has been highly research-oriented, with a handful of nice clients benefiting from my knowledge. I help my clients accelerate growth through search and content marketing strategies, new website development, and increasingly, helping my clients understand, from a practical standpoint, where and where not to apply AI to workflows, internal processes and research.
I have a particular interest in creating MCP servers to connect LLMs to external applications and data sources, harnessing these connections in processes, using my agents to crystalise procedure.
What I’m Working On Now
Most recently, I built a networked LLM server that runs Qwen 30b Instruct, with tool use via MCPO over my network.
My current projects are split between practical AI implementation, web work: building sites, apps, content and defining content strategy.
- A Shopify / SAP B1 implementation with a large hardware business located in Australia
- A SAAS application that produces data-enriched product feeds for large affiliate websites
- Automation for trends and data collection on a self-hosted instance of N8N (I’ll be rebuilding this as a small saas)
- My own affiliate content sites that drove some £15m in partner sales in 2025
- An experimental, but so far successful AI-driven job content scraper that produces xml jobs feeds from any company site by using AI to spot the patterns in new HTML before scraping, enhancing with AI and producing a jobs XML feed on Cloudflare Workers
What This Blog Is About
This blog is built for the doers: the content marketers, SEOs, vibe coders, and web workers who are ready to move past basic prompting and start building real AI processes or even infrastructure. I focus heavily on Model Context Protocol (MCP) implementation, offering step-by-step guides on setting up AI Assistant tools like Claude Desktop to interact with the real world through file operations, API connectivity and UI-less, API/MCP only apps.
From automating workflows to mastering ‘RAG without RAG’ for content research, my tutorials are designed to bridge the gap between AI theory and your daily to-do list, helping you build a smarter, more efficient workspace without wasting a huge amount of time experimenting or in R&D.
Whatever your background, I’d love to work with you to implement and execute with AI in your in-house digital processes.
Open Source & Community
I maintain a handful of open source tools at github.com/houtini-ai, primarily focused on MCP server implementations and workflow automation. If you’re building similar infrastructure or experimenting with local LLMs, the repos document what’s worked (and what hasn’t) in production environments.
Consulting Focus
I work with early-stage companies and established businesses looking to:
- Improve everyday working processes by enabling people to use practical AI tools in the workplace
- Understand and implement data strategy for Google’s vision of the orchestration layer, where commerce will take place on the AI surface, not the website
- Implement AI infrastructure for content production, specifically: the research aspect for content marketing
- Build MCP-based workflows that connect LLMs to real business data
- Content analysis or workflows for AI search engines (GEO strategy)
- Local LLM infrastructure for privacy-sensitive or high-volume tasks
- Scale technical content production whilst maintaining quality and specificity
- SEO / Content Strategy / Affiliate Merchants
Experience
Over the years I’ve naturally developed a very strong emphasis on growth marketing. In terms of recent skills and experience:
- SEO + Content Marketing
- Shopify / WordPress
- Cloudflare Workers applications, and in particular AI Workers
- Search Engine Optimisation
- N8N workflows
- Local LLM implementation on the network via LM Studio (server), MCPO (local) and OpenWebUI (local)
Contact
Want to connect? I’m on LinkedIn here: https://www.linkedin.com/in/richard–baxter or just contact me if there’s something I can help you with.
If you’re working on something interesting, have a particular problem you’d like to solve or have an operational focus that might be improved by some degree of automation, I’d be delighted to hear from you. I’d love to hear from people who might be looking for help understanding how AI might be useful in their very unique business scenarios, looking for training, looking for custom MCP servers or might already be experimenting with local LLMs, vibe coding and content creation.