The Inception of MSP-1

Where It Started

That question came from a practical place. As a photographer and software development hobbyist, I was exploring how to make my own website more accessible to language models and answer engines. I was not trying to build a protocol at first. I was trying to solve a real-world problem on a real website.

But in the process, something bigger became obvious.

The web was never truly designed to communicate clearly with AI.

The Problem Beneath the Problem

Humans can read a page and usually understand what it is, who it is for, what it is trying to say, and how seriously it should be taken. Language models can often get there too—but only after spending inference effort to reconstruct meaning that could have been declared more directly.

That gap is where MSP-1 was born.

The Core Idea

What started as an attempt to make one photography website more legible to AI evolved into a broader idea: the web needed a lightweight semantic handshake for language models. A way for content to declare intent, context, provenance, and interpretive framing up front, so machines would not have to guess so much before doing useful work.

From Personal Experiment to Open Protocol

From there, the idea grew beyond a personal implementation and into a protocol.

MSP-1 became a structured effort to create an open, machine-readable declaration layer for the agentic web, one designed not to replace content, but to make content easier for AI systems to interpret efficiently and consistently.

A Timely Way to Build

Its origin as a one-person initiative is not a weakness. It is part of what makes the project timely. MSP-1 reflects a new development reality: one human lead, working with a team of frontier language models, can now explore, pressure-test, and build serious infrastructure ideas in public.

That is part of the story too.

What MSP-1 Represents

MSP-1 is a protocol for language models that was shaped through collaboration with language models, guided by human intent, practical experimentation, and a growing realization that the future web needs a clearer way to declare meaning.

It began with one site.

It grew because the problem was much larger than one site.

About The Founder

Mark L. Johnson is a photographer and web developer with over thirty years of experience across portrait, fashion, event, and fine art photography. Known primarily for his work behind the lens, Mark has long maintained a parallel presence in web development and digital marketing, not as a technologist by title, but as a builder by instinct. MSP-1 grew naturally from that intersection: a practical problem on his own photography website that revealed a much larger gap in how the web communicates with machines. The "Mark" in Mark Semantic Protocol stands for markup, as in the declarative language of the web itself, though Mark will admit there may be just a little coincidence at play there too. For him, MSP-1 is as much about impact as it is about infrastructure. By giving AI systems clearer signals to work from, the protocol reduces the computational overhead required for inference, and with it, the energy cost. In a small but meaningful way, a more machine-readable web is also a more sustainable one. For a creator who has spent decades finding meaning in light and composition, MSP-1 represents a different kind of craft: one built for a web that is only beginning to understand itself.