mcp.clinic

• 

Inspiration

Healthcare information is fragmented. Conversations, clinical notes, lab results and messages are scattered across systems, making it difficult for clinicians to see the full context of care and for patients to understand their medical history.

We explore a simple idea: healthcare information should behave like a timeline of events, not a collection of disconnected files.

⸻

What it does

The platform converts unstructured medical inputs into a shared health timeline.

AI agents analyze doctor–patient conversations, medical notes and patient data, extract key clinical events and organize them chronologically. Instead of searching through documents, clinicians can immediately see what happened before, during and after a visit.

Patients gain a clearer view of their health journey.

⸻

How we built it

The prototype combines several components:

AI agents extract medical events from conversations and clinical notes using large language models.

Speech-to-text converts conversations into text that can be analyzed.

A timeline data model organizes events chronologically, allowing each interaction to extend the patient’s care history.

The backend processes healthcare data and feeds a web interface where the evolving timeline can be explored.

⸻

Challenges

Medical conversations are complex and context-dependent. Extracting meaningful events from incomplete statements, medical shorthand and implicit context required careful design of the event and timeline models.

Another challenge was keeping the system understandable for both clinicians and patients.

⸻

Accomplishments

We built a working prototype showing how fragmented healthcare information can be transformed into a structured timeline of care.

Instead of isolated records, healthcare data becomes a sequence of meaningful clinical events.

⸻

What we learned

Much of the real clinical context lives inside conversations and notes. Extracting events is only part of the challenge — presenting them in a form that supports medical decision-making is equally important.

The timeline model proved to be a powerful abstraction for connecting different healthcare data sources.

⸻

What’s next for mcp.clinic

Next steps include: • improving event extraction accuracy • integrating additional healthcare data sources • enabling collaboration between multiple AI agents across care teams

Our long-term goal is a platform where medical context grows with every interaction, helping clinicians coordinate care and helping patients understand their health story.

Built With

• agents
• data
• healthcare
• llm
• model
• natural-language-processing
• openai
• processing
• python
• speech-to-text
• timeline
• web