Privacy Policy

This Policy explains how the unifiedAI public website and signed-in web workspace handle personal data. It covers account, calendar, task, file and AI features made available through this website. It does not replace the privacy notices of an external service you choose to connect, and it does not describe an independently operated native app unless that app links to this Policy.

1. Who is responsible for your data?

The data controller is the operator of the unifiedAI service, presented on this website under the name “unifiedAI”. The controller decides why and how the website processes personal data. For a privacy request or question, use the contact page. When you submit the contact form, your name, contact email, organisation or sector and request are saved to unifiedAI’s support inbox for administrator review. Bug reports are saved to the same protected inbox. Submitting these forms does not open your email app.

Where a separate organisation gives you access to a workspace and decides how it is used, that organisation may also be responsible for some of your data. Contact that organisation for its instructions and privacy information.

2. The information we handle

• Account and sign-in information: your email address, display name and account settings; a salted password hash rather than your readable password; or passkey credential information. The website does not receive the biometric data or device PIN your passkey provider may use to unlock a passkey.
• Workspace information: calendar events, tasks, reminders, saved resource links, study profiles, preferences, saved response settings, conversations, messages, AI responses and files you add to the signed-in workspace. This can include personal data you choose to write or upload.
• Local study information: Local Study Mode stores study records, notes, uploaded files, flashcards, practice results and timer state in this browser’s storage. This local copy is separate from account records and can be accessible to another person using the same browser profile.
• Vault and Quick Share: the vault stores encrypted entries plus the salt and encryption metadata needed to unlock them. Quick Share stores a transfer’s name, format, encrypted content and expiry so someone with its code can retrieve it.
• AI request context: your current request and relevant recent messages, along with any file, saved profile, calendar event, task or reminder that you attach or have enabled for that feature. If you choose to share an approximate location for a reply, the assistant may receive coordinates rounded to about one kilometre and your selected time zone.
• Security records: when you successfully sign in, the workspace records the IP address and country-level information supplied by the hosting platform, plus the time of sign-in. The workspace does not use those sign-in records to record a precise street location.
• Support and reports: information you include in a bug report or email to us, such as a contact address, issue description, app version or device details. Do not include passwords, access codes or private calendar links.
• Connection and device information: the website and its hosting/security providers may process information needed to deliver pages, prevent abuse and keep the service available, such as an IP address, browser/device request details and request time. Abuse-rate counters store a salted hash derived from the request address rather than the raw address; successful sign-in history is a separate record that may contain the IP address.

You can use public information pages without creating a workspace account. We do not ask you to provide sensitive information for ordinary browsing. If you choose to put sensitive or confidential information into a prompt, file, calendar or support message, that information is handled as described here and may be sent to an external provider when needed for the feature you requested.

3. Why we use information and our lawful bases

We use information only for purposes connected with operating the features described in this Policy:

• To create and run your account and workspace: account details, settings and workspace content are used to provide the service you request and perform our agreement with you.
• To answer AI requests: the prompt and relevant context are processed to provide the response or action you ask for. Optional calendar, task and personalization context is included only when the relevant setting or feature is enabled or you supply it in the request.
• To protect accounts and the service: sign-in and connection information is used to prevent unauthorised access, rate-limit abuse, investigate incidents and keep the service reliable. Our lawful basis is our legitimate interest in protecting users and the service, balanced against your rights.
• To respond to you: contact and report information is used to deal with your request, provide support and keep a record where reasonably needed to resolve a problem. We rely on our legitimate interests in answering enquiries and maintaining the service, or a legal obligation where applicable.
• To meet legal duties and resolve disputes: relevant information may be retained or used where necessary to comply with law, establish or defend legal claims, or protect people’s rights.
• For optional browser analytics: with your consent, selecting “Allow analytics” lets the site send page paths, visit times, device categories, country codes and referrer information to its own analytics service. The server keeps only the referring website’s hostname, not its full URL. A random browser identifier, renewed after 30 days, is stored locally and converted into a keyed hash on the server to estimate unique browsers. The analytics database does not store names, email addresses, raw IP addresses, page query strings, workspace content or button labels. The request still passes through hosting and abuse-prevention systems that can process connection information. The keyed browser identifier is pseudonymous, not a guarantee of anonymity. Administrators see aggregate reports, and unique-browser counts are estimates rather than verified counts of people. Reports cover the most recent 90 days; older visit records are removed during subsequent analytics collection and administrator reporting.

Where processing is based on consent, you can withdraw it at any time. For the optional analytics choice, choose “Essential only” from the website footer’s analytics preferences or clear the site’s browser storage to reset the saved choice. Withdrawing consent stops future optional analytics collection and removes the saved analytics browser identifier. It does not automatically erase visit records already collected or affect processing already carried out lawfully before withdrawal.

Account sign-in details are needed to use the signed-in workspace. You do not have to allow analytics or provide optional AI context to browse the public pages. If you do not provide information required for a particular feature, that feature may be unavailable.

4. AI features and external providers

AI requests are processed by the provider configured for the feature or selected in the workspace. The website currently connects to Google Gemini and OpenAI. A feature may also use Google Search grounding or OpenAI web search. Some new, simple assistant requests may be retried with OpenAI if Gemini fails and fallback is configured. As a result, a request can be sent to more than one provider in that situation. We keep provider API credentials on the server.

Local Study Mode does not send its whole local database to an AI provider automatically. If you choose an AI action, the prompt, selected file and relevant study information for that action are sent to the provider. For a request, the provider may receive the text you submit, the relevant recent conversation, system instructions needed to operate the feature, and any included file or enabled calendar/task/profile context. If you use a web-search feature, the provider may also receive the query and context needed to produce search results. A provider processes this information under its own service terms and privacy documentation; its retention and handling are not controlled by this Policy.

Our OpenAI Responses API requests set response storage to off. That setting does not disable OpenAI’s separate default abuse-monitoring logs: OpenAI currently says those logs may include prompts and responses and are generally retained for up to 30 days, subject to stated exceptions. OpenAI says API data is not used to train or improve its models unless the API customer opts in. Read OpenAI’s current API data controls.

Google’s Gemini API documentation says its abuse-monitoring logs may include prompts, context and responses and are normally kept for up to 55 days, subject to exceptions. When Google Search grounding is used, Google says it stores prompts, supplied context and generated output for 30 days to create results and suggestions. Google’s terms also impose service, billing, territory and age restrictions; for example, its current terms say API clients offered in the UK must use Paid Services and must not be directed to or likely accessed by people under 18. Provider settings and terms can change, so review Google’s Gemini API terms, Gemini API abuse-monitoring information and Gemini API data-retention information before submitting information you would not want that provider to process. Do not use a feature if you do not meet its provider’s current conditions.

AI responses can be wrong or incomplete. Do not treat them as professional advice or as confirmation that a real-world change has been made. Check important information in the original source or service.

5. When information is shared

We disclose information only as needed for the feature, purpose or legal requirement involved:

• Hosting, database and file-storage providers host the website, account database and files and help secure and deliver the service.
• AI and search providers receive the request material described above when you use an AI or search feature.
• Calendar providers and apps receive information you choose to export or subscribe to. For example, adding your calendar feed to Apple Calendar gives Apple the subscription URL and event information needed to display it.
• Quick Share recipients can retrieve a transfer without signing in if they have its code, until access expires or you revoke it. This is a code-based sharing feature, not end-to-end encryption: the server encrypts and decrypts transfers to provide retrieval.
• Google’s favicon service receives requests for saved-resource website icons, including the saved website’s domain and normal connection information such as your IP address and browser details. It does not receive the saved bookmark’s title, notes or full path from the favicon URL.
• Professional advisers, authorities or a successor operator may receive relevant information where required by law, needed to protect rights or safety, or necessary for a genuine transfer of the service. We limit information to what is reasonably required for that purpose.

We do not sell personal data. The site does not use advertising cookies or send optional analytics events to an external analytics provider. Its hosting and database services process the first-party analytics described above.

6. Cookies and browser storage

The website uses storage needed to provide and protect features. This includes a secure sign-in cookie, short-lived cookies for certain account and passkey steps, and browser storage for choices such as the analytics preference, theme or device layout. Sign-in cookies are marked HttpOnly, Secure and SameSite so browser scripts cannot read them and they are sent only over secure connections to the site.

Optional analytics are off unless you choose to allow them in the privacy banner. The analytics choice enables the first-party visit records described above. Do Not Track and Global Privacy Control signals disable collection. You can choose “Essential only”, or clear this website’s browser storage in your browser settings. Local Study Mode also uses persistent browser storage for records and files; closing a tab does not normally delete it. Browser-only demonstration activity can use session storage and is separate from server analytics. Blocking essential storage may prevent sign-in or some workspace features from working.

7. Calendar links and integrations

If you create an iCalendar subscription link, it is a private access link: anyone who obtains it may be able to view the events exposed by that feed without signing into your workspace. Keep it private. You can rotate or revoke it in workspace settings; rotating or revoking prevents new authorised retrievals with the previous link. Feed responses can be cached privately for five minutes. Calendar apps may retain copies already fetched, and revocation does not delete those copies.

When you connect or export information to another provider, that provider handles the information under its own terms. Check the permissions and privacy information for that provider. A third-party app may retain a copy after it has received an export, even if you later remove the source item from unifiedAI.

Encrypted vault: vault entries are encrypted and decrypted in your browser using your separate vault passphrase. The service stores the encrypted copy and associated metadata, not the passphrase. Ordinary workspace content, AI requests and Quick Share use different handling; they do not inherit the vault’s client-side encryption. Keep an independent copy of information you need: a forgotten vault passphrase cannot be recovered by the service.

8. How long information is kept

• Account and workspace content is kept while your account is active or until you delete the relevant item or account.
• Successful sign-in records shown in security controls cover the most recent 90 days. Older records are removed during subsequent sign-ins, security-history requests and account export. Deletion can therefore occur after the 90-day point if those features have not been used.
• Session access expires after 30 days unless revoked sooner, for example when you sign out or change your password. Expiry stops access; it does not necessarily remove the database row immediately.
• Optional analytics visit records are stored on the server, not just in the current tab. Reports cover 90 days and older records are removed during collection and reporting. The browser identifier is renewed after 30 days or removed when you choose “Essential only”.
• Local Study Mode records and files persist in this browser until you delete them in Study Mode or clear the website’s browser data. Signing out or deleting a server account does not clear this separate local copy. Browser storage can also be cleared by the browser or device, so export a backup if needed.
• Quick Share transfers stop being retrievable after 10 minutes, or earlier if revoked. Expired database records are removed when an expired code is requested or a new transfer triggers cleanup; physical deletion is not guaranteed at the exact expiry time. A recipient can retain a copy already retrieved.
• Encrypted vault entries are kept in the account until you delete the vault or account. Downloaded files, copied secrets, exports and backups on your device are separate copies you control.
• Administrator access uses a short-lived session held in the current tab’s memory and checked against a hashed server record. It expires after one hour or 15 minutes without administrator interaction. Sign-in and lock events record an action and time, without the PIN, session token or IP address. They are available for 30 days; older events are removed during administrator sign-in, sign-out or security-history access. Expired session records are removed during later administrator sign-ins.
• Support, legal and security records are kept for the time reasonably needed to address the request, meet an applicable record-keeping duty, investigate misuse or resolve a dispute.
• External provider records are subject to the retention periods and controls described by that provider. Deleting a unifiedAI conversation does not itself delete a copy already retained by an AI, search, calendar or email provider.

When you delete your account, the workspace removes its account-owned records, private calendar feed and stored files from active account records and invalidates your sessions. Information that must be kept for a legal or security purpose is limited to that purpose and is deleted when no longer needed. A deletion request to unifiedAI cannot control copies already held independently by a provider or another recipient.

9. International processing

Our hosting, storage, AI, search and calendar providers may process information in countries outside the United Kingdom, depending on their infrastructure and your chosen feature. Where UK data-protection law requires a safeguard for a restricted international transfer, the applicable safeguard must be used. Contact us through the contact page if you want information about the safeguards relevant to a particular provider or feature.

10. Security

We use technical and organisational measures intended to protect account information, including salted password hashing, passkey-based sign-in where available, secure session cookies, account separation and access controls. No internet transmission or storage system can be guaranteed completely secure. Keep your sign-in details private, use a device lock, sign out on shared devices and tell us promptly through the contact page if you suspect unauthorised access.

11. Your rights and choices

Depending on the circumstances and applicable law, you may ask to access, correct, erase or restrict the use of your personal data; object to processing based on legitimate interests; receive certain information in a portable format; or withdraw consent. You may also complain to the UK Information Commissioner’s Office. These rights are subject to legal conditions and exceptions.

You can edit or delete workspace items, remove conversations, change your password, revoke calendar links and request a downloadable account export in workspace settings. For any other request, use the contact page. Local Study Mode has its own export and “Delete all local data” controls; account deletion does not operate those browser controls. We may ask for proportionate information to verify your identity before acting. Making a privacy request does not affect your statutory right to complain to the Information Commissioner’s Office.

Your right to object: you can object to processing based on legitimate interests by contacting us through the contact page. Tell us which processing concerns you so we can assess your request and explain the outcome.

12. Children and automated decisions

Some third-party AI features have their own age, location and account eligibility requirements. Those provider restrictions apply even if you can use other parts of the website. A parent or guardian’s permission does not override a provider’s age restriction. Do not use an AI feature unless you meet the eligibility conditions shown for that feature and in its provider’s current terms.

The workspace may generate AI suggestions and content, but the website is not designed to make solely automated decisions about you that produce legal effects or similarly significant effects. A person should review any AI output before using it to make an important decision.

13. Changes and contact

We may update this Policy to reflect changes to the website, providers or legal requirements. The date at the top shows when it was last revised. If a change materially affects how account information is used, we will take reasonable steps to bring it to users’ attention where appropriate.

For privacy questions, rights requests or concerns about a feature, contact us through the unifiedAI contact page. Please include the email address associated with your account and enough detail to identify the request, but never send a password, sign-in code or private calendar link.