How Lucidio Works
Simple on the outside. Thoughtful logic underneath.
Lucidio combines deterministic calculations, contextual AI generation, and a focused interface to turn complicated health information into simple context and useful questions.
Lab result pipeline
What happens between typing a result and getting questions you can bring to an appointment.
- STEP 1Enter
Choose a lab test and enter your result and the reference range printed on your report.
- STEP 2Validate
Lucidio checks that the values are complete and mathematically valid.
- STEP 3Calculate
Application logic determines whether the result is below, within, or above the user's entered range.
- STEP 4Understand
AI receives the specific test, value, unit, reference range, and calculated status to create a short plain-English explanation.
- STEP 5Prepare
Lucidio generates a few concise questions the user can discuss with a healthcare professional.
- STEP 6Save
Users can save or print the session for later appointment preparation.
The AI does not decide whether a lab result is above or below the entered range. That classification is calculated directly by the application.
One real example
Exactly what a user sees, step by step.
- Lab ResultThe user opens the Lab Result Translator.
- Hemoglobin 10 g/dLThey enter the test name, value, unit, and the 12–16 g/dL range from their report.
- Below RangeCode — not AI — calculates that 10 is below the 12–16 reference range.
- Simple explanation appearsAI writes a short, plain-English explanation of what hemoglobin is and what below range can mean.
- Questions appearA few concise questions to bring to a healthcare professional are generated.
- Save summaryThe user saves or prints the session for their appointment.
Ask Me ✨ pipeline
A short path from what you are worried about to what you can ask.
- User concern
- Context validation
- Context-aware AI generation
- 3–4 simple questions
- Healthcare conversation
Ask Me is intentionally not designed as a general medical chatbot. Its job is to turn the user's specific concern into concise questions they can bring to a healthcare professional.
What makes Lucidio different
Deterministic Where It Matters
Range classification is calculated using code rather than guessed by AI.
AI Where It Helps
AI is used for flexible plain-language explanations and contextual question generation.
Designed for Simplicity
Generated responses are intentionally short and easy to scan.
Built Around the Conversation
Lucidio does not try to replace a healthcare professional. It helps users prepare better questions for one.
Technical foundation
Only what Lucidio actually uses today. There is no proprietary medical model, no clinical database, and no server-side medical record.
Frontend framework
React 19 with TanStack Start and TanStack Router, built with Vite and styled with Tailwind CSS and shadcn/ui components.
AI integration
Google Gemini, called through a managed AI gateway inside TanStack server functions. Your inputs travel over an encrypted connection, the API key stays on the server, and nothing is stored unless you choose to save it.
State management
Local React component state for forms and results, with a unified session store that works for both guests and signed-in users.
Range calculation
Below / within / above status and the distance from the nearest boundary are computed in TypeScript, then passed to the model as fixed truth.
Validation
Client-side form checks for missing, non-numeric, or inverted ranges, plus Zod schema validation on every server function request.
Saved-session storage
Guests: sessions are stored as JSON in your browser's localStorage and never uploaded. Signed-in users: sessions sync to your account over encrypted connections so you can access them on any device.
Print output
A dedicated print stylesheet renders the Lucidio Summary as a clean document with navigation and controls hidden.
Responsive design & PWA
Mobile-first layouts from 320px upward, safe-area handling, and a web app manifest with icons so Lucidio can be installed to a home screen.
Accessibility
Semantic landmarks, labelled form fields, aria-live status updates, visible focus states, keyboard-accessible menus and dialogs, and reduced-motion support.
Error handling
Failed AI requests surface a clear error message instead of hard-coded questions, and urgent phrasing is caught before any AI request is made.
Safety by design
- Lucidio does not diagnose conditions.
- Lucidio does not prescribe treatment or medication.
- Range classification uses the user's own entered reference range.
- AI-generated language is educational and designed for appointment preparation.
- Potentially urgent concerns trigger safety messaging instead of normal generation.
For education only. Lucidio does not diagnose, treat, or replace a doctor. Lab results should always be interpreted by a healthcare professional using the patient's symptoms, history, exam, and full lab report.
Trusted information
Public health resources Lucidio links to for further reading. These links are references only — they do not imply endorsement, partnership, or affiliation.