MCP Could Make Your Health Context Portable to the AI You Choose

The following is a guest article by Bryan Barash, Founder of Wellness Project LLC

Most people who care about their health have accidentally built a pretty complicated data stack.

Your watch knows how you slept. Your workout app knows what you lifted. Your nutrition app knows what you ate. Your scale knows what you weigh. Apple Health or Health Connect may have years of activity and heart-rate data. Labs and medical records live somewhere else.

We have become very good at collecting health data. We are still surprisingly bad at letting people do much with the whole history.

That is why I think the Model Context Protocol, or MCP, could matter in health technology.

Most health apps are still built around dashboards. Open one for sleep, another for workouts, another for nutrition. Each may do a good job with its own slice of the picture.

But people do not live in slices.

The questions I actually want to ask are things like: What changed over the last six months? Why has my recovery been worse lately? Did my sleep change when I increased training volume? What was happening with my weight, food, and activity when I felt my best?

Those are hard questions because the answer may cross several systems and months or years of history. AI is a natural interface for that. You should not need to know which chart holds the answer. You should be able to ask the question in plain English.

The hard part is giving the AI reliable, permissioned access to the right information. That is where MCP gets interesting.

MCP is an open protocol that lets AI systems discover and use external tools and data sources in a consistent way. In health technology, that could mean an authorized assistant can retrieve workout history, inspect sleep trends, look at nutrition data, or pull measurements the user has made available.

The health platform can remain responsible for storing, normalizing, and securing the underlying record. The user can choose which AI is allowed to interact with it.

That is the bet we are making at Wellness Project: the health record should outlast any one app or AI assistant, while the user stays in control of who can access it.

That matters because the AI layer will change much faster than a person’s health history. Someone may prefer one assistant today and another two years from now. Ten years of health data should not become less useful because the chatbot market changed.

MCP also does not replace healthcare interoperability standards such as HL7 FHIR. FHIR helps systems represent and exchange healthcare information. MCP sits at a different layer: it gives an AI a standard way to discover what tools and data it is allowed to use.

None of this works if the security model is sloppy. Health data requires clear authentication, understandable permissions, data minimization, and an easy way to revoke access. Read access and write access should be treated differently. An outside AI provider also brings its own privacy, retention, and security practices. MCP is a connection standard, not a magic privacy shield.

For years, consumer health technology has competed on collection: more sensors, more metrics, more dashboards.

The next useful step may be simpler. Make the history we already have easier to understand.

Instead of asking, “What was my sleep score last night?” we could ask, “What changed before my sleep started getting worse?”

That is a much more interesting relationship with health data.

The most valuable health platform may not be the one with the best proprietary chatbot. It may be the one that keeps a trustworthy, longitudinal record of the person and lets that person securely bring the context to whatever intelligence they choose.

About Bryan Barash

Bryan Barash is the founder of Wellness Project LLC. He is a senior marketing and legal executive at a technology company in one of the most heavily regulated consumer categories in the country, where he works at the intersection of growth, policy, and law. Before that, he spent eight years as General Counsel to Massachusetts State Senate President Harriette Chandler after receiving his law degree from Boston College. He brings deep regulatory and legal experience, has designed and shipped websites and software himself, and is today one of the leading strategic voices on how AI is reshaping go-to-market. Wellness Project is where those threads meet: a consumer health product built almost entirely by an AI team, directed by a founder who knows regulation, code, and how to bring a product to market. Bryan lives in Beverly, Massachusetts, with his wife Claudia, son Elias, and their dachshund Seven.

The following is a guest article by Bryan Barash, Founder of Wellness Project LLC

Most people who care about their health have accidentally built a pretty complicated data stack.

Your watch knows how you slept. Your workout app knows what you lifted. Your nutrition app knows what you ate. Your scale knows what you weigh. Apple Health or Health Connect may have years of activity and heart-rate data. Labs and medical records live somewhere else.

We have become very good at collecting health data. We are still surprisingly bad at letting people do much with the whole history.

That is why I think the Model Context Protocol, or MCP, could matter in health technology.

Most health apps are still built around dashboards. Open one for sleep, another for workouts, another for nutrition. Each may do a good job with its own slice of the picture.

But people do not live in slices.

The questions I actually want to ask are things like: What changed over the last six months? Why has my recovery been worse lately? Did my sleep change when I increased training volume? What was happening with my weight, food, and activity when I felt my best?

Those are hard questions because the answer may cross several systems and months or years of history. AI is a natural interface for that. You should not need to know which chart holds the answer. You should be able to ask the question in plain English.

The hard part is giving the AI reliable, permissioned access to the right information. That is where MCP gets interesting.

MCP is an open protocol that lets AI systems discover and use external tools and data sources in a consistent way. In health technology, that could mean an authorized assistant can retrieve workout history, inspect sleep trends, look at nutrition data, or pull measurements the user has made available.

The health platform can remain responsible for storing, normalizing, and securing the underlying record. The user can choose which AI is allowed to interact with it.

That is the bet we are making at Wellness Project: the health record should outlast any one app or AI assistant, while the user stays in control of who can access it.

That matters because the AI layer will change much faster than a person’s health history. Someone may prefer one assistant today and another two years from now. Ten years of health data should not become less useful because the chatbot market changed.

MCP also does not replace healthcare interoperability standards such as HL7 FHIR. FHIR helps systems represent and exchange healthcare information. MCP sits at a different layer: it gives an AI a standard way to discover what tools and data it is allowed to use.

None of this works if the security model is sloppy. Health data requires clear authentication, understandable permissions, data minimization, and an easy way to revoke access. Read access and write access should be treated differently. An outside AI provider also brings its own privacy, retention, and security practices. MCP is a connection standard, not a magic privacy shield.

For years, consumer health technology has competed on collection: more sensors, more metrics, more dashboards.

The next useful step may be simpler. Make the history we already have easier to understand.

Instead of asking, “What was my sleep score last night?” we could ask, “What changed before my sleep started getting worse?”

That is a much more interesting relationship with health data.

The most valuable health platform may not be the one with the best proprietary chatbot. It may be the one that keeps a trustworthy, longitudinal record of the person and lets that person securely bring the context to whatever intelligence they choose.

About Bryan Barash

Bryan Barash is the founder of Wellness Project LLC. He is a senior marketing and legal executive at a technology company in one of the most heavily regulated consumer categories in the country, where he works at the intersection of growth, policy, and law. Before that, he spent eight years as General Counsel to Massachusetts State Senate President Harriette Chandler after receiving his law degree from Boston College. He brings deep regulatory and legal experience, has designed and shipped websites and software himself, and is today one of the leading strategic voices on how AI is reshaping go-to-market. Wellness Project is where those threads meet: a consumer health product built almost entirely by an AI team, directed by a founder who knows regulation, code, and how to bring a product to market. Bryan lives in Beverly, Massachusetts, with his wife Claudia, son Elias, and their dachshund Seven.