Skip to content

Security

  • Our main database, sign-in system and file storage are hosted by Supabase in London (UK, eu-west-2).
  • Storefronts and the dashboard are served through Cloudflare’s network, which also handles TLS certificates for your custom domain.
  • The sub-processors we use are listed in Data processing.
  • Everything travels over HTTPS (TLS).
  • Data is encrypted at rest by our hosting provider.
  • Secrets that should never be read back are stored only as one-way fingerprints: API keys, team invitation links, kitchen pairing codes, order tracking links, staff PINs and gift card codes. Nobody at direct.eat can read them back.
  • Webhook signing secrets and integration credentials (for example Square or Xero sign-ins) are kept in an encrypted vault and never sent to a browser.

Every restaurant’s data sits in the same database, separated by row-level security: rules inside the database itself that check, on every read and write, which organisation and shop the person belongs to and what their role allows.

  • Every table has these rules switched on.
  • Roles are enforced by the database too, not just by hiding buttons. A manager who types a web address for a billing page is refused by the database.
  • We test this with automated checks that try to read and change one restaurant’s data while signed in as another restaurant’s owner, admin, manager, staff, kitchen device, driver and diner, and as an anonymous visitor.
  • The public API follows the same rules: a key from one organisation can’t see another’s shops, orders or customers.

See Roles and permissions for what each role can do.

  • Everyone gets their own login. Team invitations are single-use, tied to one email address and expire after 7 days.
  • Sensitive changes ask for your password again: making someone an admin, transferring ownership and transferring a shop.
  • Removing someone from your team ends their access straight away.
  • Shared kitchen tablets use PIN profiles. After 5 wrong PINs, the profile is locked for 5 minutes.
  • API keys can be limited to certain shops, scopes, an expiry date and an IP allow-list.

Important actions are written to an audit log that can’t be edited or deleted through the app: menu publishes, order rejections and cancellations, refunds, team changes, exports, billing changes, app installs, integration changes and support sessions. Each entry records who did it, when, and what changed.

Team changes are shown in Settings → Team → Activity.

direct.eat staff can’t see your shop’s data by default. When you ask for help, a support agent can open a support session:

  • They must give a reason, and the session is time-limited (it ends automatically, and can be extended only once).
  • It’s read-only by default. In a read-only session, your customers’ email addresses, phone numbers and street addresses are masked.
  • To make changes, the agent has to ask. You see the request in your dashboard and can Allow for 1 hour or Decline.
  • While a session is running, your dashboard shows a banner with who is viewing, when it ends and an End now button.
  • The owner is emailed when a session starts and ends, with a summary of any changes.
  • Every session and every action in it is in the audit log.
  • Support can draft marketing offers for you, but can never approve them. See Approvals.

In an emergency (for example orders not reaching any kitchen), a senior team member can open a write session without your grant, but only with a second senior person approving it, and it’s logged as an incident.

  • Card details are entered only into Stripe’s secure payment fields on the checkout. They never touch direct.eat’s servers or database.
  • We store only what Stripe returns: card brand, last 4 digits and expiry.
  • Apps and plugins can never run on checkout or payment pages.
  • Your Stripe account is yours; we never hold your customers’ money.
  • Every webhook is signed (HMAC-SHA256) with a secret only you hold. See Webhooks.
  • Webhook addresses must be public HTTPS URLs. We refuse private networks and IP addresses, check again before each delivery, and never follow redirects.
  • API keys are shown once and stored only as a fingerprint.

If you think you’ve found a security problem in direct.eat, email security@direct.eat. Please include what you found, how to reproduce it and any proof of concept. Please don’t access other people’s data, disrupt the service or publish the issue before we’ve fixed it. We’ll acknowledge your report and keep you updated.

Was this page helpful?