Privacy Policy

Last updated: 5 October 2026.

This policy explains what Research Eagle collects, why, and what we do not do with it. It is written to match how the product actually works: the limits and the lists below are read from the settings the site itself runs on.

Who we are

This website and the Research Eagle pre-submission analysis service are operated by Research Eagle Company.

Research Eagle Company is the data controller for the personal data described in this policy. If you need to reach a person about anything here, including a request about your own data, write to hello@researcheagle.org. Messages are read by someone who can act on them.

Our postal address is not published here. If you need it for a formal or legal request, ask at the address above and we will provide it.

Your manuscript

This is the part that matters most, so it comes first.

• Manuscripts you upload are treated as unpublished confidential research.
• We do not keep your manuscript. The file is read to run the check and deleted in the same request. It is never written to our database, so there is no stored copy for you, or for us, to delete later.
• We do not share your manuscript with any publisher or journal.
• We do not make your manuscript public, and it is never placed where a search engine could find it.
• We do not train AI models, on your manuscript or on anything else.

Where the file goes. Your browser sends the file to this site, which is hosted by Hostinger. The site reads it into memory, deletes its own temporary copy, and passes it to our check engine, which is hosted by Render. The engine writes it to a temporary file, reads it, and deletes that file before it answers, whether or not the file could be read. If the engine is ever stopped in the middle of a check, it removes whatever was left behind when it starts again.

What comes back, and what is kept. The result is a list of counts and findings: word counts, the reference style, which requirements are met. A finding may quote the sentence it is about. The quote goes to your browser with the result and is stored nowhere. If you are signed in and your history is switched on, we keep one line for the check: the journal, its ISSN, the date, the verdict, the readiness score, how many items came out each way, and whether AI assistance was switched on. Not the file, not its name, not its text.

The look inside your manuscript, the corrected copy and the submission pack. After a check, the page can ask the engine to assess the manuscript against itself, to correct mechanical slips in a Word file, or to prepare an anonymous copy with a title page. Each time, your browser sends the file again and it is handled in the same way: read in a folder of its own, which is deleted before the answer is sent. The corrected copy or the pack comes back in that answer and is not kept. For a submission pack you type the authors, affiliations and ORCID iDs for the title page. They travel with the file and are not kept either.

For 30 minutes after a check, the site holds a ticket so that only the browser that ran the check can ask for these. The ticket records the kind of file, which checks were reported, the account or browser it was issued to, and how many times it has been used. Nothing from the manuscript.

When an AI model is used, and when it is not

The standard check uses no AI. Word counts, keyword counts, reference-style parsing, figure format and resolution checks, declaration detection, and the comparison of your manuscript with a journal’s requirements are all computed by rules. None of that sends anything to a language model. The same is true of the look inside your manuscript, the corrected copy and the submission pack.

AI assistance is optional and separate. It needs an account, you start it yourself, and it uses one provider: Claude, made by Anthropic. Before each request the page names the provider and lists what would be sent, and nothing is sent until you tick the box for that request.

There are four actions: “Fix flagged items”, “Research DNA”, “Journal fit score” and “Cover letter draft”. You can also chat with Claude about the manuscript you have just checked, or about academic writing in general.

AI assistance is metered. Each account includes 3 AI actions and 10 messages a month at no charge, and a paid plan adds more. The Pricing page has the details.

What is sent. Never the file. For each request you choose one of two ways:

• Send the parts the action needs. Depending on the action: the title; the abstract (up to its first 600 words); the keywords; the section headings; the first 200 words of the introduction, methods, results, discussion and conclusion; up to six references; the counts (words, references, figures, tables, and which declarations were found); what the check found; the journal’s requirements on record.
• Send nothing from my manuscript. Claude receives the journal’s requirements that the check found and what you type yourself, up to 6,000 characters.

Where the excerpt is kept. When you run a check with AI assistance switched on, the engine prepares that excerpt, signs it, and hands it to your browser. It stays in the page you have open, not on our servers, and it stops working after three hours. If you ask for an AI action, your browser sends it back and the engine passes it to Claude.

What we keep. Neither the excerpt nor the answer. For each request the engine writes one line to its log: the action, the model, the size of the request and of the answer, the time it took and what it cost. The request is counted against your allowance.

What Anthropic does with it. Anthropic processes the request to produce the answer. It states that it does not use what is sent through its API to train its models unless a customer chooses to allow that, and that it deletes API inputs and outputs within 30 days, with the exceptions it lists, such as content flagged for breaking its usage policy, which it may keep for longer. These are Anthropic’s statements, and its own pages are the record of them: Is my data used for model training? and How long do you store my organization’s data? (read on 2 October 2026).

Chat. A chat about your manuscript sends the same excerpt once, at the start, together with your messages. If you chose to send nothing from the manuscript, it sends only the journal’s requirements and what you type. A chat about academic writing sends only what you type. The conversation is held in your browser and sent again with each message. We do not store it, and it ends when you leave the page.

The site guide. The small panel in the corner of each page answers questions about this site, and it too is answered by Claude. What is sent is the question you type, the conversation so far, the address of the page you are on, and the page’s language; when you ask about a journal, also what our catalogue holds for it. It carries no manuscript and no account details, and it needs no account. We do not store the conversation. Each browser may ask 12 questions an hour, and each IP address has a limit of its own. The cookie and the counters behind those limits are described under Cookies.

Contact and enquiry data

If you contact us through the form on this site we collect your name, your email address and anything else you choose to include, so that we can reply. The form sends them to us by email. The site keeps no separate record of form entries, but it does keep a log of the email it sends, holding the recipient, the subject and the body of each message, for 30 days. We keep your message in our mailbox while the enquiry is open and for a reasonable period afterwards for our own records.

The form includes an automated spam check: a hidden field that only a script fills in. It inspects the submission itself and does not build a profile of you.

Accounts

If you create an account, we keep:

• your name, your email address, a username made from that address, and the date you agreed to this policy;
• your password, in hashed form only. You set it through a link we email to you;
• what you choose to add to your profile: institution, research area, ORCID iD, a short bio and a photo. A photo you replace or take off is deleted;
• your preferences: which kind of check is selected when the form opens, the language for Claude’s answers, and whether to keep a history;
• how much of your AI allowance you have used today and this month, and any credit you have bought, with the date it expires;
• unless you switch it off, a history of your last 30 checks: the journal, the date and the result, never the file or its text. You can clear it at any time on your account page;
• while you are signed in, a record of the session: when it began, when it ends, and the IP address and browser it was started from. WordPress, the software this site runs on, keeps this so that a session can be ended, and removes it once the session is over.

We use these to sign you in, to show you your own account, and to email you about it. We never sell them, and we never share them with publishers.

If you sign in with Google or ORCID, that service tells us who you are. We store the account ID it gives us, so that we recognise you next time, and your name and email address if it passes them on. Where it does not pass on an email address we ask you for one and confirm it with a link. We never see your password for that service.

If you change your email address, the new one is held as pending until you open the link we send to it, for up to a day.

You can delete your account at any time from your account page. That removes your profile, your photo, your preferences, your history and your allowance from the site’s database. Payment records are kept, as described next.

Payments

Paid plans are handled by Wayl (wayl.io), a payment provider. You enter your card, wallet or phone details on Wayl’s own page, never on this site, and Wayl processes them under its own privacy policy. A plan is a single payment. It does not renew, so nothing is kept for charging you again.

To start a payment we send Wayl the plan, the amount and a reference of our own. We do not send your name or your email address. (A reference made before 3 October 2026 contains your account number.) The price is shown in US dollars and charged in Iraqi dinars; your bank converts it.

From Wayl we receive whether the payment went through, the amount and the payment method (for example “Card”). We keep a record of each payment: the plan, the amount, the date, its status, the method, Wayl’s reference, your account number, and the technical details that let us confirm Wayl’s notice is genuine. We use it to add the plan to your account and for our accounts.

Within 14 days of paying, if nothing the plan added has been used, you can ask for a refund from your account page. We record when you asked and when it was refunded, the refund is made through Wayl, against the same payment, and the plan’s credit comes off your account. A charge made twice, or a payment whose plan never reached your account, is refunded whatever the date: write to us.

On your account page you can also take your name off your payment records (“Remove me”). The plan, the amount, the date and Wayl’s reference stay for our accounts, without your account number, and we record when this was done. Deleting your account does the same. Wayl keeps its own record of your payment.

Checks are free and, with an account, unlimited. What a plan adds is AI actions, chat messages and submission packs, and we keep count of how many you have left and when they expire.

Cookies, and what is kept in your browser

We do not use advertising cookies, and we do not use third-party tracking or profiling cookies. These are the cookies this site sets:

• Signing in. WordPress sets its own cookies when you log in, so that the site knows you are signed in. They last until you close your browser, or for 14 days if you tick “Keep me signed in”. While you are signed in, the site’s cache sets one more, so that a page made for you is never shown to anyone else. Three short-lived cookies are used only while you are signing in or signing up: rea_return (30 minutes) remembers which page to bring you back to, and reso_bind (15 minutes) and reso_pending (30 minutes) are used when you sign in with Google or ORCID.
• The free check without an account. rec_guest, a random number that lasts 30 days. It lets the site count your free check for the day without knowing who you are.
• The site guide. reg_v, a random number that lasts a day, set when you first ask the guide a question. It is how your questions for the hour are counted.
• Language. rea_lang, which lasts a year and remembers whether you chose Arabic or English on the check and account pages.
• Statistics. _koko_analytics_pages_viewed, which lasts until midnight and lets our own page counter count one visitor once. See Statistics below.

To make its free limits work, the site also keeps a few counters on its server: how many checks, searches, sign-ups or guide questions came from one browser or one IP address. Each counter is a number. It is filed under a keyed hash of the address, never the address itself, and it expires within a day.

In your browser’s own storage, the check page remembers the last journal you chose, so that you do not have to type it again. It stays on your device.

Statistics

We count page views with Koko Analytics, which runs on this site and sends nothing to anyone else. It records which pages are viewed and which sites visitors came from, as daily totals, and keeps them for 36 months. It does not record IP addresses and does not build a profile of a visitor. Visits by administrators and editors are not counted.

We also count, per day, how many checks were run with and without an account, how many sign-ups, AI actions, messages, submission packs and corrected copies there were, how many plans were paid for (by plan), and how many checks were run against each journal (by the journal’s ISSN). A count is only a number: it holds no IP address, cookie, account or anything you typed, so it cannot be traced to you. Staff are not counted. The counts are kept for 36 months.

Who else is involved

We do not sell your personal data, and we do not share it for advertising. These are the other companies involved in running the service, and what each one handles:

• Hostinger hosts this site and its database, delivers its pages, and carries the email it sends.
• Render runs the check engine. It handles a manuscript for the seconds a check takes.
• Anthropic provides Claude. It receives what is described above, and only when you ask for AI assistance or ask the site guide a question.
• Wayl takes payments.
• Google and ORCID, if you choose to sign in with one of them.
• Google also serves the typefaces this site uses (Google Fonts). Your browser fetches them from Google’s servers when a page loads, so Google receives your IP address and the address of this site.

The site and its database are hosted in Frankfurt, Germany, and the check engine runs in Frankfurt, Germany. The other companies above process data where they operate.

How long we keep things

• Your manuscript: not kept. It is deleted in the request that reads it.
• The excerpt for AI assistance: in your browser, for up to three hours.
• The ticket that follows a check: 30 minutes.
• Counters for the free limits: a day at most.
• Cookies: as listed above. The longest is the language choice, a year.
• Your account, profile, preferences and history: until you delete them, or the account.
• Payment records: kept for our accounts. After “Remove me”, or once an account is deleted, they are kept without the account number.
• Email the site sends: logged for 30 days.
• Enquiries: while the enquiry is open and for a reasonable period afterwards.
• Statistics: 36 months, as totals.
• Technical logs: like any web server, our hosts record each request (the time, the address requested and the IP address it came from) and keep those logs for their own limited periods. What you type into a form or the journal box is sent in the body of the request, not in its address. The check engine’s own log records each request by its route, its outcome and the time it took, and never the contents of a manuscript or anything you typed.
• Backups: Hostinger, which hosts the site, makes a copy of the whole site, its database included, every day, and stores it in France, apart from the server in Frankfurt. Hostinger deletes old copies on its own schedule and does not state the period for this plan; on 5 October 2026 the oldest copy it held was from 6 September 2026. Something you delete can remain in a backup made earlier until that backup expires.

Your rights

Depending on where you live, you may have the right to ask for a copy of your data, to have it corrected or deleted, to restrict or object to how it is processed, and to receive it in a portable form. You can also complain to your local data protection authority.

To make any of these requests, write to hello@researcheagle.org.

Security

We take reasonable technical and organisational measures to protect the data we hold, and the Security page describes them. No service can promise perfect security, and we will not claim otherwise.

Changes to this policy

If this policy changes materially we will update the date at the top, and where the change is significant we will tell users directly rather than relying on you to notice.

Contact

Questions about this policy, or a request about your data, can be sent to hello@researcheagle.org. You can also use the contact form.