Privacy
Last updated 10 November 2026.
This policy covers the Rick app for iPhone and Mac, and the pages people publish from it on this domain. Rick is made and operated by Egg Matter Ltd, a company registered in England and Wales under number 17505486, with its registered office at 5 Brayford Square, London E1 0SG, United Kingdom. Egg Matter Ltd is the data controller for the processing described here. Questions and requests: [email protected].
The short version
- Your notes live on your device. If you turn on sync, they are also kept in your own iCloud account. We can't read either copy.
- There is no sign-up. We never ask for your name, email address or password.
- Rick uses random IDs to tell devices and iCloud accounts apart. They contain nothing about who you are, but they do tie records to one device or account.
- Some features send data to our server and to the companies that do the work, such as AI and transcripts. Each one is described below.
- There is no advertising and no third-party analytics. We don't track you across other companies' apps or websites, and we never sell data.
Your notes
Your notes are stored in a database on your device. Backups you export are files in your control.
If you turn on iCloud sync, which comes with the Sync and Pro plans, your library is also copied to your own private iCloud database so your other devices can read it. That includes your notes, attachments and their text, link previews, images and PDFs, flashcards and reviews, tags, assistant conversations, the records of pages you have published, and your settings, including a first name if you gave one. Each change also carries the name of the device that made it, such as "Sam's iPhone". Apple holds that copy for you, and we cannot read it.
Locking a note hides it in the app. It does not add encryption beyond what iCloud already applies, so a locked note syncs like any other. Your own AI keys, if you add any, and the check for your note-lock code travel between your devices in iCloud Keychain.
Our server does not store your notes. The exceptions are pages you choose to publish and the search vectors described below.
The IDs Rick uses
Device ID. The first time Rick needs one, it makes a random ID for that installation. It is sent with most requests to our server.
Account ID. When your device is signed in to iCloud, Rick links it to your iCloud account each time it opens, whether or not you use sync. It sends our server a fingerprint (a one-way hash) of your iCloud account's identifier, and a short-lived token that our server uses to check the sign-in with Apple. Our server keeps only a keyed fingerprint of that identifier. It gives your account its own random ID, and records which devices belong to it and when each was last seen. This lets your plan, purchases and AI allowance follow you across your devices.
Neither ID contains your name, email address or Apple ID, and we can't look up who you are from them. Because they tie records to one device or account, the App Store lists this data as "Linked to You".
What leaves your device, and why
The assistant
Before anything goes to an AI company, Rick asks for your permission and names the companies involved.
When you use the assistant with Rick's AI, your message, the conversation so far, the note you are working in, and any other parts of your notes the assistant looks up are sent through our server to Anthropic, whose Claude models write the answer. Under Anthropic's API terms, this content is not used to train their models. Our server passes it on and does not keep it.
The assistant can search the web using Anthropic's search tool. The search words it chooses are sent to Anthropic.
When your message includes links, our server asks a small Anthropic model how to handle them, so your message and the links are sent to Anthropic for that step. Our server may also fetch those pages or their transcripts for the assistant.
On the Pro plan you can add your own key for Anthropic, OpenAI or DeepSeek. The assistant's requests then go straight from your device to that company, under your own account and its terms, and our server doesn't see them. The link step, page and transcript fetching, and the assistant's searches of your library still go through our server.
For each AI request we keep a usage record: your device and account IDs, the time, which part of the app made the request, the model, the amount of text processed, the cost, a random conversation ID and how long it took. It never includes what you or the assistant wrote. We use these records to run allowances, the usage gauge and fair-use limits.
Search by meaning
Once you have agreed to AI, the text of your notes and attachments is sent through our server to Voyage AI, which turns it into vectors: lists of numbers that let the assistant find notes by meaning, not only by matching words. The assistant's searches of your library, and the start of anything it adds to a note (to suggest tags), are turned into vectors the same way. The vectors are stored on your device. Our server also keeps a copy of each vector, filed under a fingerprint of the text rather than under your IDs, so the same text isn't processed twice. Locked notes are never sent for this, and neither are files that sit only in locked notes. Until you agree to AI, nothing is sent for this. The search field in Rick always matches words, so it works either way.
The free trial
The first time you use the assistant on the free plan, Rick asks Apple's DeviceCheck service whether this device has had the trial before. Apple keeps a "trial used" mark for the device, which stays even if you delete and reinstall Rick. Our server records the trial credit against your IDs.
Checking requests come from Rick
Rick uses Apple's App Attest so our server can tell that requests come from a genuine copy of the app. Our server keeps the public key for your installation, a counter, and when it was last used. On a Mac where App Attest isn't available, our server checks your iCloud sign-in with Apple instead and gives that Mac a pass that lasts a few hours.
Link previews
When you add a link to a note, our server fetches the page to get its title, image and similar details. For some sites, including Facebook, Reddit, Threads, Medium, Quora, Google Docs and some Instagram links, your device fetches the page itself and sends the page's code to our server to read. Our server may also ask public services such as YouTube and Spotify for details of a link. If our server can't reach YouTube, your device may ask YouTube directly. On the Mac, Rick also reads page titles directly from the sites.
Images and site icons in previews load directly from the sites that host them. YouTube thumbnails come through our server, which fetches them from YouTube. Some site icons come from Google's icon service, which receives the site's address.
On the Mac, if you switch to Rick while a web link is on your clipboard, Rick sends that link to our server to fetch its title, so the title is ready if you paste it. The Mac may also fetch the page itself. This happens even if you don't paste. Links that look like sign-in, password-reset or unsubscribe links are skipped.
Our server keeps the pages and transcripts it fetches, filed by the link and not by who asked for them, so the same link loads faster next time.
Transcripts
When you ask for the transcript of a video or podcast, our server gets it from YouTube through our transcript service, or downloads the episode's published audio and has it transcribed by Groq or Deepgram. To find a podcast's audio, our server looks the episode up on Spotify, YouTube, Apple Podcasts and the show's own feed. This is always the show's published audio, never a recording of you. Our server keeps your IDs on each transcription job until it is finished, then removes them within a day. The usage record for the request stays, like any AI request's.
Purchases
Apple handles payment, and we never see your card details. RevenueCat manages subscriptions and top-ups for us. Rick identifies you to RevenueCat by your account ID, or your device ID before your device is linked. RevenueCat's code also sends its own device details: on iPhone, the identifier Apple gives each app maker for your device, plus the device model, system version, language and App Store country. We keep your plan, its renewal and expiry dates, the transaction numbers of top-ups, the credit each one added, and any refunds.
Walkthrough steps
The walkthrough you see the first time you open Rick sends us which of its steps you reach and which of its actions you complete, with your device ID and the time. It never includes what you write. This shows us where the walkthrough confuses people.
Sync health reports
While sync is on, Rick checks once a day, and whenever you run Check Sync, that your iCloud copy matches your device. It then sends us a report: how many items of each type are on the device and in iCloud, how many differ, how many changes are waiting, any errors, a short technical message if the check failed, the platform, the app version and your device ID. It never includes the content of your notes. This lets us spot sync problems before they cost anyone their work.
Pages you publish
Publishing a note with Share to Web uploads a page for each note you include, with its images, GIFs, PDFs and their preview images, and a preview image for link sharing. Cloudflare hosts the pages for us, and anyone with the link can read them. A page can show thumbnails from other websites; a visitor's browser loads those directly from the site that hosts them.
For each published page, our server keeps a random page ID, a fingerprint of the secret that lets you take the page down, the number of pages, the file names with a fingerprint of each file's contents, your account or device ID, when it was published and taken down, and who took it down. The secret itself syncs with your library so any of your devices can take the page down. When a page comes down, its files are deleted. The record and the file fingerprints stay as our audit trail for handling reports and for recognising the same content if it is published again.
Reports
If you report a page through the report form, we keep the page's link, the category, your description, your email address, whether you agreed to be identified to the author, a keyed fingerprint of your IP address, and the dates the report was received, confirmed, acted on and replied to, with its status and outcome. Report emails, including the confirmation link and our reply, are sent through Resend.
Your description is shown to the page's author, and anyone who has the page's link can also see it. Your email address is not shown, unless you agreed to that for a defamation complaint. If someone repeatedly misuses the form, we keep a fingerprint of their email address so we can recognise further reports from them. If you email [email protected] instead, we keep the correspondence for the same reason we keep reports: to show how a complaint was handled.
Server logs
Heroku, which runs our server, keeps a short rolling log of requests. It shows each request's IP address, time and path, which can include a link you added or your device ID. Heroku keeps only the most recent lines, and we use them only to fix problems. Our server also holds IP addresses in memory for about fifteen minutes, to limit how many requests one address can make. We don't store IP addresses anywhere else, apart from the fingerprint on a report.
Why we use your data
- To provide what you ask for: the assistant, search by meaning, link previews, transcripts, purchases, the trial and published pages.
- In our legitimate interests: checking requests come from a genuine copy of Rick, preventing abuse, keeping sync working through health reports, and improving the walkthrough. These use only IDs, counts and steps, never the content of your notes.
- To meet legal duties: keeping company and tax records, and handling reports about published pages.
Who else handles data
| Company | What they receive |
|---|---|
| Anthropic | Assistant requests, the link step and web searches |
| Voyage AI | Note and attachment text, to make search vectors |
| Groq, Deepgram | Published podcast audio, to transcribe it |
| Vercel, Webshare | Host our YouTube transcript service and its connection to YouTube; they see the video's link |
| Video and channel IDs, episode titles to find a matching video, and site addresses for icons. Email sent to [email protected] is delivered to our Gmail inbox. | |
| Spotify, Apple Podcasts and other sites' public services | Links and episode details, for previews and transcripts |
| Apple | Payments, iCloud sync and Keychain, the iCloud sign-in check, App Attest and DeviceCheck |
| RevenueCat | Subscriptions and top-ups |
| Cloudflare | Hosts published pages and forwards email sent to [email protected] |
| Resend | Sends report emails |
| Heroku | Runs our server and database, and keeps short logs |
| OpenAI, DeepSeek | Only if you add your own key for them; under your own account |
Some of these companies are outside the UK, mostly in the United States. Transfers rely on the safeguards that UK law approves, set out in their terms.
How long we keep it
| Data | Kept for |
|---|---|
| Walkthrough steps | 12 months |
| Sync health reports | 90 days |
| AI usage records | 24 months. Records that count towards a trial, a beta allowance or a top-up are kept for as long as that balance needs them. |
| Search vectors on our server | 90 days |
| Fetched pages, transcripts and link details | Kept, filed by link and not by person, so the same link loads faster |
| Your IDs on a transcription request | Until the job is finished, then removed within a day |
| Purchase and credit records | At least 6 years, as company records. A top-up balance never expires. |
| Account ID, device links and App Attest keys | While you use Rick, or until you ask us to delete them |
| Published-page records and file fingerprints | As long as we may need to show how reports were handled |
| Reports | As long as we may need to show how a complaint was handled. The email address and its fingerprint are removed after 2 years, and the IP fingerprint after 1 year. |
| Server logs | A short rolling window, usually days |
What we don't do
No advertising, no third-party analytics or tracking code, no tracking across other companies' apps or websites, no selling or sharing data for marketing, and no access to your notes on your device or in your iCloud. If a future feature changes any of this, this page will say so before the feature ships.
Age
Rick is for people aged 18 and over. We don't knowingly collect data from anyone younger. If you think someone under 18 has used Rick, email [email protected] and we will delete what we hold about them.
Your rights
UK data-protection law gives you rights over the personal data we hold: to see it, correct it, have it deleted, object to how we use it, and complain to the Information Commissioner's Office (our registration number is ZC274886). To use them, email [email protected]. Because Rick has no accounts with names, we may ask for details that help us find your records.
Changes
This policy may change. The date at the top tells you when it last did.