Privacy Policy

This Privacy Policy explains how AGGRAI LTD ("aggrai", "we", "us") collects, uses, and shares information when you use aggrai (the "Service"). We are a company registered in England and Wales with company number 17180867, and our registered office is at 123 Buckingham Palace Road, London SW1W 9SH, United Kingdom. We are the controller of your personal data. By using the Service you agree to the practices described here.

1. Information we collect

We collect the following categories of data:

• Account data: when you create an account, your email address and a securely hashed password (handled by our auth provider). If you sign in with Google or GitHub instead, we receive the basic profile details they share with us, such as your name, email address and profile picture. We also keep the language you use aggrai in with your account: the one you choose, or, if you haven't chosen one, the one you signed up in or your browser's. We use it to write our emails to you in that language, to show aggrai in a language you chose when you sign in on another device, and to ask Stripe to use it for your receipts.
• Questions and selections: the questions you ask, any images or PDFs you attach, the AI models you choose, and the resulting responses (including any images you generate). We also work out which language each question is written in and store that with its result, so that answers display correctly (for example, right to left for Arabic) and so that we can see how many questions are asked in each language.
• Conversations & personalisation: if you're signed in and continue a comparison, we save the conversation thread — your follow-up questions and the answers — so you can return to it across devices. You can also set an optional personalisation profile (how you like answers written — length, tone, background, interests). If you switch on “Fill this in for me” (off unless you turn it on), we additionally store short facts aggrai infers from the questions you have asked. All of this is per-account and can be reviewed, edited or cleared in Settings → Personalisation.
• Plan and billing: the tier you're on (Free, Standard, or Premium) and any related transactional records.
• Technical & analytics data: only with your consent, a pseudonymous per-browser id, your IP address, coarse device/browser type and your browser's language setting — used to understand usage, catch errors and decide which languages to offer. If you reject analytics we do not store these, and any IP address we do store (with consent) is erased within 90 days. Separately, we always process an approximate country from your connection to choose a language variant and, for web searches, which country to search (see Brave Search in section 4), under our legitimate interest (see “If you reject analytics” below).
• Cookies & local storage: strictly-necessary cookies to keep you signed in, to remember your cookie choice and, if you choose one, to remember your language, plus optional analytics — error monitoring and the pseudonymous usage identifiers above — that run only with your consent (see the table below). We do not use advertising or cross-site tracking cookies.

If you reject analytics, we switch off the items marked “optional” above and clear their identifiers. We still process what we genuinely need to run aggrai: your account (to sign you in and provide your history), the questions you ask and the models you run — with their token and cost accounting and the language each question is written in — and a coarse country to choose a language variant and a web-search region. Those records carry no per-visitor id, IP address, device profile or browser language, and we rely on our legitimate interest in operating and securing the Service (and, for account data, on performing our contract with you). Signing in is unaffected by this choice.

2. How we use it

• To provide and operate the Service (routing your question to the models you selected, returning their answers).
• To tailor answers to you, by including the personalisation preferences and any learned facts you have chosen to store (Settings → Personalisation) in the prompt sent to the model. This applies to every question you ask while signed in, including the first one in a conversation. It is never applied to the cached example questions on the landing page (those answers are shared between all visitors), and it is never used to train models.
• To enforce usage limits and prevent abuse. Whatever your cookie choice, we use your IP address to limit how often questions can be asked, conversations shared or copied, the contact form sent, and analytics events sent. We never store the address itself: it is first turned into a scrambled code using a secret key. For questions, that code is held only in memory; for the other limits it is kept in a short-lived counter that is deleted within about a day. If you reach a limit, the code may appear in our security logs, which are kept for up to 90 days. We rely on our legitimate interest in protecting the Service for this.
• To make sure each person gets only one free trial from a promotional code. When you start one, we keep a scrambled code made from your email address with a secret key (it can be matched, but not turned back into your address), your card's fingerprint (an identifier Stripe gives each card, not the card number), and which code and subscription the trial was for. We check them whenever a free-months code is used, and rely on our legitimate interest in preventing misuse of our offers for this.
• To improve the Service — for example by caching responses to common questions so they load instantly for everyone, and by logging how our internal question classifier categorised your input so we can tune it over time.
• To respond to support requests submitted via our contact form (stored with your name, email and message, and deleted after 90 days).
• To respond to support requests and security incidents.
• To comply with legal obligations.

We do not sell your personal data and we do not use your personal questions to train our own models.

Note: your recent comparisons are also kept in your browser tab's session storage so you can switch between them instantly. That data never leaves your device and is cleared when you close the tab.

3. Third-party AI providers

When you submit a question, the question text, any files you attach and your model selection are forwarded to the third-party AI providers you chose, via the OpenRouter routing service. We also send questions and answers through OpenRouter to models we choose (normally Anthropic's Claude) to classify each question, write the summary and scores, tag its topic and, if you have switched on “Fill this in for me”, pick out facts for your personalisation. Each provider processes that data under its own privacy policy. Avoid submitting confidential or personally identifying information in your questions or files.

Current providers may include:

• Anthropic (Claude) — privacy policy
• OpenAI (GPT) — privacy policy
• Google (Gemini) — privacy policy
• Meta (Llama) — use policy
• Mistral — terms
• OpenRouter — privacy policy

4. Service providers we use

We rely on the trusted providers below to run the Service. They process data on our behalf only as required to deliver their service (if you choose to sign in with Google or GitHub, their own privacy policies also apply to that sign-in):

• Supabase — account authentication and database.
• Vercel — application hosting and edge runtime.
• Fly.io — hosts our API server, which every question and answer passes through.
• Cloudflare — DNS, and storage for the images and PDFs you upload and the images you generate.
• Sentry — error monitoring. Before an error report is sent, we strip out IP addresses, cookies, sign-in tokens and what you typed, such as your questions. In your browser, Sentry runs only if you Accept analytics.
• OpenRouter — AI model routing (see section 3).
• Brave Search— web search, when a question needs up-to-date information. It receives a search query based on your question (which we may reword or translate) and a country to search in. We choose that country from your question — the language it is written in, or the place it is about — or, if the question doesn't point to one, from the approximate country of your connection. Brave never receives your account details or your IP address.
• Stripe — payments, subscriptions, tax and receipts. Your card details go into a secure Stripe form, so we never see your full card number. We also tell Stripe the language you use aggrai in, so that its receipts and emails to you are in that language where Stripe supports it.
• Resend — sends our emails, such as account, billing and contact-form emails.
• Microsoft 365 — our hello@aggrai.com inbox, where contact-form messages and emails you send us arrive.
• Google — optional sign-in with your Google account.
• GitHub — optional sign-in with your GitHub account, and our scheduled background jobs, such as tagging questions by topic and updating personalisation for people who have switched it on.
• Slack — internal alerts and a daily activity summary for our team. The summary contains only counts and totals, never email addresses or questions.
• Anthropic — an AI assistant (Claude) that investigates errors for us. It can see the question and the AI output from a comparison that failed.

5. Caching of common questions

To keep the Service fast and affordable we cache the responses to common example questions and reuse them across users. Cached responses are not tied to your account and do not include any personal identifiers.

6. Data retention

• Account data is retained for as long as your account is active. On account deletion it is removed within 30 days, except where we must retain it to comply with legal obligations, or as described below.
• Cached example responses are retained for up to 24 hours and refreshed periodically.
• Saved conversations are retained until you delete your account, and your personalisation profile until you clear it or delete your account. Both are included in your data export and erased when you delete your account.
• Files you upload are kept until you delete your account. Images you generate are deleted automatically after 30 days.
• Comparisons you share by link stay publicly viewable, and deleting your account does not remove links you have shared. A shared link also records the language you were using aggrai in when you shared it, so that its preview on other sites is written in that language.
• Messages you send through our contact form are deleted from our database after 90 days. A copy also goes to our support inbox, with the language you were using aggrai in so that we can reply in it, where we keep it only as long as we need it to deal with your request.
• If you start a free trial with a promotional code, the record described in section 2 (the scrambled email code, card fingerprint, code and subscription) is kept until 12 months after the trial ends, including if you delete your account in that time, when it is unlinked from your account. It is used only to stop another free trial.
• Operational logs (errors, abuse signals) and any IP address stored with your consent are retained for up to 90 days, after which IP addresses are automatically erased.

7. Security

We use reasonable technical and organisational measures to protect your data: HTTPS in transit, password hashing via our auth provider, access controls, and regular updates. No system is perfectly secure; we cannot guarantee absolute security.

8. Your rights

Depending on your jurisdiction you may have the right to access, correct, export, or delete the personal data we hold about you, and to object to or restrict certain processing. You can download a copy of your data or permanently delete your account yourself at any time from Settings → Privacy & data. To exercise these rights another way, email hello@aggrai.com. We may need to verify your identity before responding.

9. International transfers

Our infrastructure providers operate globally; your data may be transferred to and processed in countries other than your own, including the United States and the European Union. Where required, transfers rely on standard contractual clauses or equivalent safeguards.

10. Children

The Service is not intended for children under 18. We do not knowingly collect data from children. If you believe a child has provided us with personal data, contact us and we will delete it.

11. Changes to this policy

We may update this policy as the Service evolves. Material changes will be announced via the Service or by email; the "Last updated" date at the top of the page always reflects the current version.

12. Contact

Questions, requests, or complaints? Email hello@aggrai.com.