Updated November 7, 2026

Security & trust

WEPAIR AI is built so you can check how it works, not just take our word for it. Here's what protects your information, and the limits of what we claim.

Your information stays yours

  • Private by default. Every table in our database has row-level security: the database itself only returns your cases, photos, appliances and notes to your account. An automated check fails our build if any table is missing it.
  • Signed-out visitors get read-only database access to public content only (like plans and the Explore posts people chose to share). An automated check fails our build if anything gives them write access.
  • Your photos and conversations are never shared with other customers or repair businesses.
  • Secret keys never reach your browser. Database admin access and payment keys exist only on our servers, and an automated check looks for leaks on every build.
  • Payments are handled by Stripe. WEPAIR AI never sees or stores your card number.

Sharing happens only when you choose

  • Private links — a technician confirmation link or an appliance passport — are long random links. They expire, you can withdraw them, and they show only what's listed when you create them. Only our server can open them.
  • An appliance passport shows the appliance and its confirmed repairs, never your account, case titles, conversations, photos or anyone's name.
  • WEPAIR staff see counts on our internal growth page, not your records.

How Ava earns trust

  • Where this came from. Answers list each source Ava used — manual, web page, your own records or another customer's reported fix — with how far to trust it. If no document was used, the answer says so.
  • Checked by plain code. After Ava answers, WEPAIR's checker removes citations to sources that weren't found and flags error-code meanings no source contains, "the manual says" with no manual, and false certainty.
  • Facts vs guesses. Records are labelled by how sure we are: manufacturer documentation, technician-confirmed, customer-reported, sensor reading or AI suggestion.
  • AI requests aren't kept by the AI provider. We ask the provider not to store each request, and an automated check confirms every AI call in our code does this.
  • Safety first. Ava never advises bypassing safety devices and tells you when to stop and call a professional.

What we don't claim yet

WEPAIR AI is young. We haven't had an independent security audit or certification yet; we'll say so here when we do. Ava's answers are AI-generated guidance — check model-specific steps in your manual and use a qualified professional for gas, mains electricity and other hazards.

Found a security problem? Please contact us through the contact page and we'll respond quickly.