Dictate, anonymise, structure — reflect.

By default, no. Speech recognition runs entirely on-device using Apple's frameworks and the audio never leaves your device. On a small number of older devices where on-device recognition isn't available, the app will tell you clearly before anything is sent — you'll see an explicit warning and have to actively agree before Apple's servers are used. If that fallback is active, you'll be reminded not to include patient-identifiable details. Either way, no audio is ever stored on our servers, and anonymisation always runs locally before any text is processed further.

An on-device scrubber identifies patient names, dates of birth, specific locations and rare diagnoses, and replaces them with generalised placeholders ([patient], [date], [tertiary centre]) before any text is transmitted. You can review and edit the scrubbed text before it's sent.

Subscriptions are managed by Apple. Open Settings → [your name] → Subscriptions on your iPhone, choose Portfol.io, and tap Cancel. Cancelling before the renewal date stops the next charge; existing entries remain accessible until the period ends.

In the app, open Profile → Privacy & data → Delete account. Your account and all stored entries are removed from our servers within 30 days. This is permanent and irreversible.

Each entry has a Copy action on every field, mirroring the structure of your ePortfolio form (CBD, mini-CEX, ACAT, etc.). Bulk PDF export is on the roadmap — email if you'd like to be notified when it lands.

Entries are mapped against UK postgraduate curricula — JRCPTB CiPs, Generic Professional Capabilities and equivalent frameworks. You can adjust the mapping per-entry before copying into your ePortfolio.

Treat it as a drafting aid, not a finished entry. You're responsible for reviewing every field before submission. Verified NICE guideline citations are linked back to source so you can confirm them.

iPhone and iPad running iOS / iPadOS 17 or later. On-device transcription is required for the privacy model to hold — if your device doesn't support it, the app will tell you explicitly on first launch. No silent fallback to cloud speech recognition.