MHRA-registered medical device

Your medical
notes, explained.

Paste a doctor's letter, test result, or discharge summary. Get a clear explanation in seconds.

Translation, not interpretation: Patiently explains what your clinician wrote — it never adds advice or assumptions. The approach →

Try it free — no login Download the app
verifiedRated 4.5/5 for medical accuracy in our validation study emoji_events5 national awards check_circleFree on web, iOS & Android

Your text isn't saved · Available on iOS, Android & web

For healthcare organisations & partners

Watch: what Patiently AI does, in two minutes

summarize

Summarise in seconds

Turn complex letters, results, and discharge summaries into clear summaries you can actually read.

translate

Explain in plain language

Reading levels from child-friendly to clinical. 18 languages. Tailored to your needs.

shield

Private & secure

Identifiers removed on your device from typed or pasted text. Nothing stored. Nothing used for AI training.

Communiqué Awards 2025 Winner Healthcare & Scientific Comms Communiqué 2025 BMA 2025 Platinum Winner Platinum Winner BMA 2025 HTN AI & Data Awards 2025 AI for Comms & Prevention HTN Awards 2025 PMEA 2025 Winner — Patient Education Patient Education PMEA 2025 PMEA 2025 Winner — Innovation Innovation PMEA 2025
01 — The problem

Millions of patients can't understand
their own health records

Medical letters are written for clinicians, not patients. The result: confusion, anxiety, and worse outcomes.

9–11
Average reading age
of the UK adult population
43%
Struggle with health information
written at clinical reading levels
10 min
Average GP appointment
not enough time to explain everything
Patient looking at medical documents
02 — What it does

Clarity in seconds, not hours

translate

Plain-language explanations

Complex medical information translated into everyday language that's easy to understand and remember.

help_outline

Questions for your doctor

Generate effective, relevant questions for your next appointment. Walk in feeling confident and informed.

language

18 languages supported

Access explanations in multiple languages, making healthcare accessible for diverse populations worldwide.

monitor_heart

Keep a health diary

Symptoms, blood pressure, mood and medications in the iOS app, kept on your phone to take to your clinician.

tune

Tailored to your level

Adjust the reading level and communication style to match your needs — from child-friendly to clinical.

group

Share with your care team

Seamlessly share important medical insights with doctors and caregivers to improve care coordination.

lab_profile

Lab results at a glance

Paste a blood test report and see your results in a table next to the explanation. Every value, range and flag is copied exactly from your report — nothing is calculated or added.

menu_book

Guides to common tests

Plain-English guides to 30+ blood and urine tests, from HbA1c to eGFR. Browse the guides

summarize

Patient-friendly summaries

Generate clear, accessible summaries instantly, reducing the time spent explaining complex terms during busy clinics.

forum

Better communication

Improve engagement with accessible, clear explanations that patients actually understand and remember.

mic

Simplify by voice

Dictate and simplify medical notes effortlessly, empowering patients to feel more in control of their health.

speed

Streamline your workflow

Automated explanations free up consultation time, letting you focus on personalised care rather than translation.

verified

MHRA-registered

Built to clinical standards as a registered Class I medical device. Safe to recommend to patients with confidence.

api

API & enterprise access

Connect via secure REST API or deploy privately within your own environment for full data governance.

03 — Use cases

Built around real patient needs

A patient receives test results full of medical jargon — and panics before they've even seen their doctor.
Patiently AI translates results into plain language, replacing anxiety with understanding before the follow-up appointment.
Understanding test results
Patient use case
A patient wants to ask the right questions at their appointment — but doesn't know where to start.
Patiently AI generates relevant, personalised questions from the clinical notes, building confidence before the consultation.
Preparing for appointments
Patient use case
A GP has 10 minutes per patient and no time to translate complex terms into everyday language.
Patiently AI generates patient-friendly summaries instantly, freeing up consultation time for personalised care.
Streamlining consultations
Clinician use case
Patient character
04 — The evidence

Scientifically validated

Rigorous research with healthcare professionals and patients confirms Patiently AI's effectiveness.

87.3%
Rated clinically safe
By expert clinician review (binary judgement)
7 in 10
Patients prefer it
Over their original clinical notes
3.0
Grade levels easier
Significant readability improvement

Validation study published in Frontiers in Digital Health, 2026 · doi: 10.3389/fdgth.2026.1771051
Design framework published in SN Comprehensive Clinical Medicine, 2026

The study used OpenAI's GPT-4o. When GPT-4o was retired, we moved to a newer model in October 2026 only after a bridging evaluation against the study: the same notes, pre-specified pass criteria and a blinded expert review. The new model kept every dose and test value and added no numbers not in the original.

05 — See it in action

How Patiently AI turns medical
jargon into clarity

Patiently AI
Clinical note
auto_awesomeSimplified
helpQuestions for your doctor
    06 — Your care, protected

    Built to clinical standards

    We hold ourselves to the same standard clinicians set when they care for patients.

    science

    Validated in practice

    Tested with real healthcare professionals and patients. The peer-reviewed validation study, published in Frontiers in Digital Health, 2026, confirms accuracy, safety, and genuine readability improvement. Patiently AI's design principle — translation, not interpretation — is the subject of a peer-reviewed paper in SN Comprehensive Clinical Medicine, 2026.

    verified

    Registered medical device

    MHRA-registered Class I medical device. Built to regulatory standards, not as an afterthought — compliance is foundational.

    shield

    Privacy without shortcuts

    When you type or paste a note, personal identifiers are removed on your device before anything is sent. Uploaded photos are read by our AI provider first, so crop out names and other details if you can. GDPR-compliant, ICO-registered, Cyber Essentials certified. Your medical information is never shared or used for training, and is not stored beyond a 24-hour retry window.

    From the validation study

    Known limitations & how we mitigate them

    Expert clinicians flagged three concern categories during peer review. Each is addressed in product — transparency over hype.

    • Flagged Over-simplification could minimise serious conditions
      → Reading-level controls (Child → Carer); every output recommends clinician follow-up.
    • Flagged Adding general advice not present in the original note
      → Design principle: translation, not interpretation — the subject of our SN Comp Clin Med paper.
    • Flagged Patient misinterpretation of simplified text
      → Built-in "Questions for your doctor" feature; standard "not medical advice" disclaimer on every output.
    • Found by our eval Simplification could drop safety-critical specifics (a dose, a test value)
      → Caught by OpenGATE, our open-source verification framework, on its first run — fixed the same day and now regression-gated in CI: every backend change is scored before it ships, and the live service is re-scored weekly. 100% of critical facts preserved, zero invented numbers. Since October 2026 every explanation is also checked as it is generated: if a dose, test value or timeframe from your note is missing, it is regenerated once. Lab tables are checked the same way — any number not in your report is removed.
    Your data is never stored Nothing used for AI training Registered medical device
    UKCA Class I ICO ZA712068 GDPR DPIA Cyber Essentials
    07 — For healthcare organisations

    Embed plain-language explanations
    in your product

    A regulated medical-text simplification layer for patient portals, EHRs, hospitals, and digital health platforms. Built by clinicians and a PhD medical writer. Validated by independent research.

    Built for
    monitor_heart

    Patient portals & EHRs

    Plain-language layer on letters, results, and discharge summaries.

    local_hospital

    Hospitals & NHS trusts

    Cut "what does this mean?" calls; improve health literacy.

    medication

    Pharma patient support

    Companion content for trials, PSPs, and patient comms.

    female

    Femtech & women's health

    Make ultrasound, screening, and gynae reports patient-friendly.

    groups

    Charities & public health

    Multilingual patient education at population scale.

    What you get
    api

    Secure REST API

    Endpoints for simplification, questions, explainers, and medication info. JSON in, JSON out. UK/EU hosting on request.

    cloud_done

    Private deployment

    Hosted in your VPC, GCP/AWS/Azure tenancy, or on-prem. Bring your own model keys or use ours.

    extension

    White-label embed

    Drop the PWA into an iframe or web view with your branding. Fastest route to "live in your product" — typically 2–4 weeks.

    language

    18 languages

    Patient-facing output in 18 languages including Italian, Spanish, French, German, Polish, and Arabic.

    5,000+medical notes simplified
    100+countries reached
    18languages supported
    MHRAClass I medical device
    Compliance & security
    MHRA Class I UK MDR 2002 GDPR ICO ZA712068 Cyber Essentials DPA available DPIA support EU residency on request
    No customer data retained by default. Text is processed in-memory and discarded after the response. The exceptions are a de-identified copy of the reply kept for up to 24 hours so a retry in the iOS app is free, notes users choose to keep in the app's History, and share links, which users explicitly create and which auto-expire after 90 days.
    Never used for training. Customer text is not used to train any AI model.
    On-device PII redaction. Identifiers (NHS, MRN, names, addresses, phones, emails) stripped from typed and pasted text before processing. Uploaded photos are sent for text recognition before redaction.
    Sub-processors disclosed. OpenAI (text and text recognition from photos), Google Firebase (hosting), ElevenLabs (TTS), Resend (email). Full list in DPA.
    Pricing model
    Pilot
    Fixed fee
    3-month engagement · request a quote
    • Scoped use case, single integration surface
    • Up to ~10k simplifications
    • DPA & DPIA support
    • Outcomes review at end
    Most flexible
    Production
    Usage-based
    Annual platform fee + volume tiers
    • REST API at production volume
    • UK/EU hosting on request
    • SLAs, support, dedicated contact
    • Co-marketing on request
    Embedded
    License
    Annual licence + integration fee
    • White-label embed in your product
    • Private deployment / your tenancy
    • Bring-your-own model keys option
    • Roadmap input
    Request a 30-min intro Request DPA & security pack
    Or email info@pharmatools.ai — we respond within 24 hours.
    08 — Questions

    Frequently asked

    For patients
    Patiently AI simplifies medical notes, letters, and test results into clear, everyday language. It helps patients understand their health records and generates questions for doctor appointments.
    Yes. When you type or paste a note, personal identifiers are removed on your device before anything is sent. If you upload a photo, it is sent to our AI provider to read the text first, so crop out names and other details if you can. PDF and Word files are read on our server, then de-identified before any AI sees them. The AI provider doesn't use your text for training. Your medical text isn't stored on our servers, apart from a de-identified copy of the reply kept for up to 24 hours so a retry is free, and any notes you choose to keep in the app's History. We're GDPR-compliant, ICO-registered and Cyber Essentials certified.
    Appointment letters, discharge summaries, test results, prescription details, referral letters, treatment plans, surgery reports, and specialist consultation notes.
    Yes. In the web app you can upload a photo (JPEG, PNG, WebP or iPhone HEIC), a PDF or a Word file instead of pasting text. When you type or paste a note, personal identifiers are removed on your device before anything is sent. If you upload a photo, it is sent to our AI provider to read the text first, so crop out names and other details if you can. PDF and Word files are read on our server, then de-identified before any AI sees them.
    Yes. Patiently AI is registered with the MHRA as a Class I medical device under UK MDR 2002, with Medcopywriter Limited as the responsible person.
    Our published validation study rated outputs 4.49/5 (mean) for medical accuracy on a 5-point clinician Likert scale, with 87.3% of outputs judged clinically safe to share with patients. 70% of patients preferred Patiently's explanations over the original clinical notes. Full study published in Frontiers in Digital Health, 2026. The study used OpenAI's GPT-4o. When GPT-4o was retired, we moved to a newer model in October 2026 only after a bridging evaluation against the study: the same notes, pre-specified pass criteria and a blinded expert review. The new model kept every dose and test value and added no numbers not in the original.
    For healthcare organisations
    Yes. Three integration paths: a secure REST API, a white-label web embed (fastest — typically 2–4 weeks), or a private deployment in your tenancy. Contact info@pharmatools.ai to scope your use case.
    The public PWA is hosted on Google Cloud (currently us-central1). By default no customer data is retained — text is processed in-memory and discarded after the response. The exceptions are a de-identified copy of the reply kept for up to 24 hours so a retry in the iOS app is free, notes users choose to keep in the app's History, and share links: if a user explicitly creates one, that note is stored in Firestore and auto-expires after 90 days. UK and EU residency are available on request for enterprise deployments, either via region-locked Google Cloud / AWS, or private deployment in your own tenancy.
    Yes. A Data Processing Agreement is available, and we can support your DPIA with sub-processor disclosure, retention policy, and security controls. Sub-processors currently include OpenAI, Google Firebase, ElevenLabs, and Resend.
    No. We don't use customer text to train any AI model, and OpenAI does not train on API data. OpenAI may keep requests for up to 30 days for abuse monitoring, then deletes them. Read-aloud audio is generated by ElevenLabs under its standard terms.
    18 patient-facing languages including English, Italian, Spanish, French, German, Polish, Arabic, Mandarin, Portuguese, Hindi, and Urdu. Additional languages can be added for enterprise customers.
    Three structures: Pilot (3-month fixed fee), Production (annual platform fee + usage-based volume tiers), or Embedded (annual licence for white-label or private deployment). Request a quote for your use case.
    A white-label embed is usually live in 2–4 weeks. A REST API integration into your product takes 4–8 weeks depending on your team's bandwidth. Private deployments are scoped per engagement.
    Character illustration

    Two ways to use Patiently AI

    Free for patients · partnership-ready for healthcare organisations.

    For patients

    Understand your health, on your terms

    Free on web, iOS and Android. No login required. Your text isn't saved.

    Try Patiently AI
    Download on the App Store Get it on Google Play
    For partners

    Embed plain-language explanations in your product

    API, white-label embed, or private deployment. MHRA Class I. DPA on request.

    Request a 30-min intro
    Patiently AI is a registered Class I medical device. It provides simplified explanations of medical text for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional for clinical decisions.
    From the PharmaTools.AI lab

    Patiently AI is built in the open by Nick Lamb, PhD — an independent AI researcher & engineer working on evaluation, verification and oversight of AI systems.

    Explore the research →More free tools