Privacy Policy
Last updated: 16 November 2026
This Privacy Policy explains how Akunudo LLC-FZ (“we”, “us”) collects, uses, keeps and shares data in connection with AlphaCitation (the “Service”). It describes what the Service does today.
1. Data we collect
- Account data: your email address and sign-in records. If you pay by card, the name and email you enter at checkout.
- Organisation data (enterprise clients): your organisation's name; the email address, role and status of each seat; market groupings; and the signed contract with its value and renewal date.
- What you enter to run scans: brand, category, competitors, alternative names, market, audience and any positioning statement. We also record brand and category text typed into our suggestion fields, even if no scan is run.
- What you record in the product: claims you register (their text, whether approved or prohibited, reference codes and supporting evidence), case notes, and the names of people you assign as owners. We record the email address of the person who made each entry.
- Your organisation's assistant connection and data API: for each tool call made with your organisation's API key or signed-in assistant connection, we record the tool or operation used, the day (and the time of each tool's first and latest use), the kind of credential (API key or signed-in connection) and the client software the call declares (for example Claude, ChatGPT or a programming library). We also count, across all callers, how many calls to each tool failed. Each read is entered in your organisation's activity log, with the time, the person's email address and role (or, for an API key, its name and first characters), the programme and the number of records read; for a search through your assistant this includes the search text (up to 200 characters) and, for each programme searched, the number of results. Your organisation's owners, administrators and reviewers can see and export the log. We do not record API keys (only a key's name and first characters, which tell keys apart) or the content returned.
- AI answers we record: for subscription scans, we store each question we put to an AI service, the full answer it returns, any sources it cites, the model and the time. See section 5.
- Public posts: to report public sentiment, we collect public posts and comments that mention a brand from YouTube, X, GitHub, Stack Exchange and Hacker News. These can include the author's public display name.
- Enquiries and report requests: if you use our enterprise enquiry form, we store your name, work email, organisation, role, the brands you list and your message. If you ask us to email you a free scan, we store your email address, the brand and category of that scan, and a browser identifier (see section 4).
- Emails we send: for each email we send, we keep the recipient, the subject and a copy of its content.
- Payment data: card payments are handled by Stripe. We do not receive or store your card number. We keep your plan, its status, Stripe's customer and subscription identifiers and, if you cancel, the cancellation reason Stripe records.
- Usage and technical data: server logs, including IP address, and the analytics described in section 4. When our open API or MCP server is called, we also record the endpoint, the time, the outcome and response time, the software the caller declares (for example a browser, a programming library or an AI assistant), the country, the website the call came from (its domain only), and a tag, made from the address under a key deleted after two days, that links one caller's calls within a single day. A call to the reference lookup also records the medicine, market, claim type and indication it names, and a call to the Evidence-Authority judge or requirements the claim type and market; a call to the MCP server records only the method and the name of the tool it calls. Its arguments travel in the request body, which no layer records, our hosting provider's server logs and the encrypted security log included. We look up which organisation owns the network a call came from. Where that network is registered to an organisation, we record the organisation, its industry and the evidence for the attribution. We never record the organisation behind a home or mobile connection. Request bodies, credentials and free text are not recorded.
- API keys: when you create an API key, we collect the organisation name and domain you provide, securely store a non-reversible identifier of the key (never the key itself), and record its usage, including when it was last used. We associate API activity with your organisation to manage access, maintain security, understand adoption and improve our services. We retain this information for as long as reasonably necessary for these purposes, including while your account remains active, subject to our data retention policy. We also keep the name you give each key, its first few characters (which tell your keys apart, and are not the key), and, where you reached the key page from one of our links, a short label naming that link (such as a launch post); the label never identifies a person. Your keys are deleted when your sign-in is deleted.
2. How we use data
- to provide the Service: run scans, and generate and deliver reports, alerts and digests;
- to process payments and manage subscriptions;
- to keep a dated record of what AI services said in your scans;
- to maintain security and prevent abuse;
- to improve the Service, including by using measurements from many scans in aggregated form;
- to understand which organisations, developers and AI services discover and use our open API, MCP server and publications, to improve them, and to follow up commercially with organisations whose use suggests an interest;
- to give your organisation a record of what was read through its assistant connection and data API, and to understand which tools each organisation uses and how often, so we can improve them, support integrations and ask organisations whose use suggests an integration whether they have built one; and
- to send prospective customers a sample report with an invitation to subscribe.
3. Service providers and other recipients
- Supabase: our main database, sign-in, and private storage for report PDFs and signed contracts.
- Vercel: hosting and server logs.
- Upstash: a data store for scan results, sessions and caches, and the queue that runs scans.
- Stripe: payments. Stripe also receives the brand, category and competitors you enter at checkout.
- Postmark: email delivery. It receives each email's content, including attached reports.
- PostHog: website analytics. On signed-in and report pages we remove sign-in codes from page addresses, do not send report addresses, and do not record clicks or screen content.
- AI services (OpenAI, Anthropic, Google Gemini and Perplexity): they receive the questions we build from your scan inputs (brand, category, competitors, market and audience). OpenAI or Anthropic also receive the content of your report and the questions you ask the report assistant. Public posts to be classified go to Anthropic, or to OpenAI or Google when Anthropic is unavailable. Each provider handles this data under its own API terms.
- Search and public-data services: Google (search suggestions, Google News and YouTube), X, GitHub, Stack Exchange and Hacker News (through Algolia) receive brand and category names as search terms. Public medicine registers, for example those of the US Food and Drug Administration and the European Medicines Agency, receive medicine names.
- FreeTSA: an independent timestamping service. It receives a cryptographic fingerprint of a record, never its content.
- IPinfo and the regional internet registries (such as ARIN and RIPE NCC): they identify the network an address belongs to. They receive the IP address of a call to our open API or MCP server, once per network each week.
We do not sell your personal data.
Scorecard pages. When a free scan is run, its summary (brand, category, market, overall score and the competitors named) is published at a web page whose address contains a random code. Anyone with the address can open it, and search engines may index it. The summary of a subscriber's scan is available in the same way to anyone who has its address, but it is marked so that search engines do not index it. Scans for enterprise clients have no public page. The full report is never on a scorecard page.
4. Cookies and similar technologies
- Sign-in and report access: Supabase sign-in cookies; ac_unlock (report access, 24 hours); ac_remember (keeps you signed in for 30 days if you choose).
- Your workspace: ac_brief_seen (when you last opened your Brief on this device, 1 year), so the Brief can mark what is new since then.
- Free-scan identifier: ac_vid, kept for one year. It lets us recognise a browser that runs another free scan, and it links that browser's free scans (brand and category) to any report request made from it.
- Analytics: PostHog's cookie and browser storage, on every page except our administration console. Session recording is limited to our public pages.
- We do not use advertising cookies.
5. Data retention
We do not currently delete data on a schedule. Specifically:
- Kept until deleted: your account and brands, your report PDFs, enquiry-form submissions, free-scan report requests, free-scan records linked to the ac_vid identifier, and claims and case records.
- Kept after a brand is deleted: the AI answers we recorded for its scans, and the scan history, measurements, citations and snapshots derived from them, and its public-sentiment history, which can include short verbatim excerpts of public posts quoted as evidence for a theme. When you delete a brand we keep these, no longer linked to your account but still naming the brand.
- Kept permanently by design: an entry in our deletion log (brand name, category, market and date) when a brand is deleted, the signed evidence records and case histories issued for it, our record of each integration an organisation has confirmed to us, and your organisation's activity log, which records each view, change and read (the database refuses any change to or deletion of its entries, including when a brand or the organisation is deleted, so it remains a reliable record).
- Kept for a limited time: stored scan results, 180 days; copies of sent emails, 30 days; collected public posts, 30 days; server logs, about one day; the encrypted security log of calls to our open API and MCP server, with IP addresses, 30 days; records of those calls without IP addresses, about 13 months; the organisation records built from them, and records of which tools each organisation used and when, until about 13 months after the organisation's last call.
6. International transfers
We operate from the United Arab Emirates. Our providers process data in other countries, including the United States and the European Union.
7. Your rights
Subject to applicable law, you may ask for access to, correction of, or deletion of your personal data, or object to or restrict certain processing. Email us at analyst@alphacitation.com. We handle deletion requests manually. Some records cannot be deleted without breaking records that other people rely on: the AI answers recorded for your scans, case histories, signed evidence records, your organisation's activity log and the brand deletion log. Where a record cannot be deleted, we tell you which records are kept and why.
8. Security
We use reasonable technical and organisational measures to protect personal data. No method of transmission or storage is completely secure, and we cannot guarantee absolute security. IP addresses of calls to our open API are stored encrypted, are readable only by our administrators, and are deleted after 30 days.
9. Children
The Service is not directed to anyone under 18, and we do not knowingly collect their data.
10. Changes
We may update this Policy; the "Last updated" date above reflects the latest version.
11. Contact
Akunudo LLC-FZ: analyst@alphacitation.com.
Questions about this document? analyst@alphacitation.com.