Skip to content
HARDEX
Back to home

Privacy policy

Last updated: November 11, 2026

Hardex checks web apps for security problems. This page explains, in plain words, what information we keep when you use it, why, and what you can ask us to do with it.

Hardex is run by AryonX.ai (“we”). The service is new, so we keep only the information the product really needs.

1What we collect

Only what is needed to run your account and your scans.

  • Your account: your email address and the name Google shares with us. You sign in with Google, so we never receive or store a password.
  • Your scans: the web address you asked us to check, when the scan ran and how long it took, your score, the technologies we recognised, and every problem we found together with the fix we suggested. A finding can quote a short, partly hidden piece of your app, for example the first characters of a key.
  • For an uploaded project, a fingerprint of its files: one value made from the files as a whole with a key that only our server has, kept with the scan. It cannot be turned back into your code or into a list of your files. It is used for one thing, to tell you when you upload exactly the same files again, so that you do not wonder why nothing changed. It is deleted with the scan.
  • Shared report links, only if you make one: which scan it shows, when it was made and until when it works. Only a fingerprint of the link is kept, never the link itself, so we cannot show it to you again.
  • Deep review, only if you turn it on for an uploaded project: when you agreed, the model used, how many files were read, how long it took and what it cost, and every problem it reported: the file and line, the kind of problem and a short note written by the AI. Never the code itself.
  • Your settings: which checks you switched on.
  • Your plan: which plan your account has and, once you pay for one, when it renews or ends. For a paid plan we keep the identifiers Polar gives your purchase (a customer number and a subscription number), the plan, how often you are billed, whether the subscription is active, and when it renews or ends. We never receive or store your card number.
  • For the deeper test that creates test records: when you agreed to it, which database it ran against, and the one-time code you placed there to prove the database is yours.
  • Operating records: our hosting providers keep short-lived records, such as the time of a request, the address it came from and any error, so the service can run and abuse can be stopped.
  • Scheduled scans, if you turn them on: the web address, how often, when the next and the last scan ran, and the email address of your account, which is where the alerts go.
  • API keys, only if you make one for your build: its name, when it was made and when it was last used. Only a fingerprint of the key is kept, so we cannot show it to you again.
  • Scan allowance record, written only when you delete your account: a one-way fingerprint of your sign-in email address, the month, and how many scans your account made that month. It exists so that deleting an account and signing up again does not give back the month's scans on your plan. The fingerprint is made with a secret key that we keep, so it cannot be turned back into your address from a copy of our data. It holds no scans, no problems and no name.

2What we do not collect

  • We run no advertising or tracking tools, and we do not sell information about you.
  • We do not keep code you upload for a scan. It is deleted the moment the scan ends, or the deep review if you asked for one, and only the results stay.
  • We do not receive your Google password, your contacts, your calendar or any other Google data. Google only tells us who you are: your name and email address.
  • We do not use your scans to train AI models. We send your code to an outside AI service only when you turn on the deep review for that project and agree on the screen that explains it.

3How we use it

  • To sign you in and keep your account safe.
  • To run your scans and show you the results, your score before and after a fix, and your history.
  • To protect the service and the sites people scan: we limit how many scans one person can run, and we may look at records to investigate abuse.
  • To find and fix faults in the service.
  • To scan again the websites you chose, on the schedule you chose, and to email you when something new turns up.
  • To run the deep review when you ask for it, and to count how many reviews you have used this month.
  • To show a scan’s report to the people you give a link to, until the link expires (1, 7 or 30 days, your choice) or you stop it. The link shows the report and the problems found; it never shows your account, your email address, your fix prompts, the raw evidence or any other scan, and anyone who has the link can open it.
  • To let your own build send a project for a check with a key you made, until you revoke the key or delete your account.

4Who else handles your information

We use a small number of services to run Hardex. Each one handles your information only to provide its part:

  • Supabase: sign-in, and the database where your account and scan results are stored.
  • Railway: runs the scanner.
  • Vercel: serves this website.
  • Google: you sign in with your Google account.
  • osv.dev, a public database of known weaknesses in software packages: when you upload a project, we send it the names and versions of the packages the project uses, to check them. Your code is not sent.
  • npm and PyPI (registry.npmjs.org, api.npmjs.org and pypi.org), the public registries where software packages are published: when you upload a project, we may ask them when each package the project uses was first published and, for npm, how often it is downloaded, to spot packages that appeared only days ago. We send them the names of the packages that come from these public registries and nothing else; they see the request come from our server, not from you. Your code and your account details are not sent, and a private package, or one from your own registry, is never asked about.
  • Anthropic: when you turn on the deep review for a project, it receives the text of that project’s code files so its AI can read them. Hardex tries to replace secret values, such as keys and passwords, with a placeholder first, but cannot promise to catch every one. Hardex sends it through Anthropic’s API and does not store the code. How long Anthropic keeps what it receives is set by its own terms.
  • Resend: sends the alert emails if you turn on scheduled scans, and the one test email if you ask for it. It receives your email address and the message, which says only the name of your website, how many new problems there are, your score and a link (the test email says only that it is a test).
  • Polar: when you buy a paid plan, Polar handles the payment as our payment provider and merchant of record. When you go to pay, we send it your email address, an internal number that identifies your account with us, and a signed code that lets us match the purchase to your account. Polar also receives your name, your billing country and your payment details, which you enter on its checkout and not on our site, and it keeps its own records of the purchase, as the law requires of it. We receive from it the confirmation of your purchase, your plan and when it renews or ends. Polar handles your information under its own privacy policy.
  • Sentry: receives a short error report if Hardex breaks, in your browser or on our server: the error message and where in our code it happened, the kind of browser you use, the address of the page or request without anything after a question mark, and, for the browser, the last few pages and requests that came before it. We have set it up so that it does not receive your scan results, the addresses you scanned, your email address, your sign-in or what you typed or clicked. Like any service your browser connects to, it can see the network address (IP) of your connection.

These companies run servers in several countries, so your information may be handled outside Israel. We do not share it with anyone else, except where the law requires us to.

5Cookies and storage on your device

Hardex stores only what it needs to work: a sign-in cookie that keeps you signed in, a cookie that remembers your language, and a small note in your browser that remembers which checks you picked and how you arranged the dashboard. There are no advertising or tracking cookies.

6How long we keep it

We keep your account until you delete it. We keep each scan for 365 days (see the next paragraph). You can delete your account yourself at any time, from Settings in the dashboard: it removes your scans, findings, deep reviews, accepted risks, automatic scans, settings and your sign-in record at once. If you pay for a plan and the subscription is still being billed, the account is not deleted until you have cancelled the subscription, so that nobody is billed for an account that is gone. You can also ask us by email, and we delete them within 30 days. The short-lived operating records are removed by our providers on their own schedule.

We do not keep scans for ever. A scan is removed 365 days after it was made, together with its findings, fixes, deep review and any shared report link for it. The newest scan of each website or project in your account is the one exception: it stays, so that your dashboard is never empty and a new scan can show what changed. It goes when you delete your account. Accepted risks and settings are not scans, and stay until you remove them or delete your account.

A schedule stays until you turn it off, use the link in any alert email that stops it, or delete your account.

One small record outlives your account, until the month it belongs to has ended: the scan allowance record described under "What we collect". We keep it to stop the limits of a plan from being reset by deleting and re-creating an account, and we remove it within a day after the month ends.

7Your choices

You can ask us to:

  • show you the information we hold about you,
  • correct anything that is wrong,
  • delete your account and everything linked to it, except the scan allowance record until its month ends (you can also do this yourself, in the dashboard).

Write to us using the contact details at the end of this page. We may need to check that the request really comes from you before we act on it.

8Keeping your information safe

We protect your information with a sign-in check on every request, separate access for each account (you can only see your own scans), encrypted connections, and limits on how the scanner can be used. No online service is perfectly safe, and we cannot promise that nothing will ever go wrong. If something affects your information, we will tell you.

Found a security problem in Hardex itself? Please tell us at the address at the end of this page.

9Who can use Hardex

Hardex is for people aged 18 or over. We do not knowingly collect information about children.

10Using Hardex responsibly

Hardex is meant for checking websites and apps that are yours, or that you have clear permission to check. Any misuse of Hardex is entirely the responsibility of the person who misuses it, as set out in the Terms of Use.

11Changes to this policy

If we change this page in a way that matters, we will update the date at the top. For a big change we will also tell you by email or in the app.

12Questions or requests

Write to us at support@aryonx.ai. We read every message.

ARYONX (AryonX.ai), Israel