Medbuddy - Making Prescriptions & Clinical Data Easier to Understand with AI Data Extraction
Overview
MedBuddy is an AI-powered tool that turns complex prescriptions and medical documents into clear, human-readable guidance that patients and healthcare professionals can understand in seconds. Built as a session-based, no-login experience, it supports multiple file types and instantly generates structured explanations covering diagnoses, medications, dosages, lab results, interactions, and safety precautions. The product is designed for people who often feel overwhelmed by medical jargon as well as clinicians who need a fast, reliable way to simplify instructions for patients. I led the full UX/UI design of the platform — from brand tone and visual direction to the desktop-first dashboard layout, microcopy, component design, edge-case handling, and the landing page that introduces the experience. The result is a clean, trustworthy interface that makes medical information more transparent, actionable, and shareable.
Introduction
Why MedBuddy Matters
Designing MedBuddy introduced several core challenges centered around clarity, trust, and variability. Medical documents come in countless formats, mixing handwritten notes, abbreviations, and inconsistent spacing, so the design needed to adapt to unpredictable inputs without overwhelming users. At the same time, the tool had to work for two very different audiences: laypeople who need plain language and reassurance, and healthcare professionals who expect precision and detailed breakdowns. With no login system, exporting and sharing also had to be effortless while maintaining strong privacy guarantees. Trust was another critical dimension — the interface needed to look authoritative but still approachable, using visual cues that support safety without feeling clinical or intimidating. Finally, the product had to gracefully handle failure states, from poor scan quality to ambiguous lab markers, while keeping the user grounded, informed, and able to continue.
Problems & Challenges
The Core Problems We Needed to Solve
MedBuddy needed to make sense of prescriptions that vary widely in format, clarity, and completeness — while serving two very different audiences: patients seeking simple explanations and clinicians expecting structured accuracy. The product also had to build trust from the first interaction, since everything happens in a session-only environment with no accounts and no data storage. Finally, the interface needed to handle edge cases gracefully: failed scans, missing information, unclear lab values, and mixed handwriting or abbreviations.
So, how might we…
…explain medical details simply without oversimplifying important clinical meaning?
…design a consistent UI for documents that are inherently inconsistent?
…communicate privacy and safety clearly in a lightweight, reassuring way?
…support error states smoothly (e.g., failed scans, incomplete data)?
…keep the workflow fast and intuitive: upload → understand → export → done?
Project Goals
One for the User, One for the Business, One for the Tech
Understand prescriptions safely and confidently
Understand prescriptions and lab results instantly, with explanations written in clear human language and actionable guidance.
Build a scalable, compliant MVP for early adoption
Demonstrate Easybits’ AI capabilities through an intuitive, trust-driven public product that showcases structured extraction and reasoning.
Standardize free-form PDFs into structured data
Support multiple document types (PDF, image, DOCX, TXT) with a consistent output structure, all in a session-based environment with no user accounts.
Solution
A Clean, Trust-First Experience
MedBuddy delivers a clean, trust-first experience built around one simple promise: upload a prescription and instantly understand what it means. The interface uses a dashboard-like layout with clear structure and a calm, clinical look that reinforces safety without feeling intimidating. Users begin by uploading a document, after which the system processes it and presents the results in modular sections—Diagnosis, Medications, Lab Results, Precautions, and Recommendations—each designed to be readable for patients but detailed enough for clinicians through a layperson/doctor toggle. Everything is organized in a predictable pattern so users never feel lost, even when the prescription itself is messy or confusing. A lightweight chat assistant gives people a safe space to ask follow-up questions without leaving the page, while the option to export a clean PDF ensures they can save or share the results since no data is stored. By combining structured outputs, calm visuals, and clear “nothing is saved” messaging, the experience feels both private and reliable. The overall flow stays intentionally short: upload → understand → export → done—resulting in a product that feels reassuring, fast, and genuinely helpful.
Results
How MedBuddy Improved Clarity, Trust, and Workflow Efficiency
The first iteration of MedBuddy demonstrated how a focused, privacy-first experience can transform how patients and clinics read prescriptions. By combining structured parsing, a clean layout, and an optional medical chat, the tool reduced confusion, increased confidence, and supported real-world clinical workflows—all without requiring accounts or storing data.
Results & impact:
Faster understanding: Users could interpret prescriptions in seconds, thanks to structured sections and plain-language explanations.
Higher trust signals: Privacy-first design, clear disclaimers, and session-only processing increased user confidence during early demos.
Clinician-ready output: Doctors valued the ability to switch to medical terminology and quickly validate medication details.
Improved export workflows: PDF exports enabled patients to share results with healthcare providers, compensating for the absence of accounts.
Future-proof structure: Modular card layouts made it easy to add new components—like Lab Results and Precautions—without redesigning the system.
Efficient engineering collaboration: A design that respected technical constraints (stateless chat, rigid output model) allowed rapid iteration with fewer blockers.
Learnings
Key Learnings
Working on MedBuddy sharpened my ability to design for a high-stakes environment where clarity, privacy, and accuracy matter more than visual flair. The core challenge was translating messy, inconsistent medical documents into a calm, guided experience that works for both patients and clinicians—without ever storing user data. This required balancing UX simplicity, technical constraints, and clinical nuance.
Key learnings:
Trust is a design outcome, not a feature. Every interaction—from upload to export—had to reinforce privacy, safety, and credibility.
Modularity scales better than complexity. Building results as independent cards allowed new medical sections (like labs) to integrate seamlessly later.
Plain language unlocks accessibility. Designing dual terminology modes showed how much clarity impacts user confidence and comprehension.
Lightweight flows reduce cognitive load. A single-page dashboard kept users focused and avoided decision fatigue, especially important for older or anxious patients.
Constraints drive better solutions. Stateless chat, no accounts, and limited parsing models pushed the design toward a simpler, faster experience that users actually prefer.
Clinical contexts reward humility. Collaborating with engineering exposed gaps in AI interpretation—reminding me that design must guide, not overpromise.