Privacy Policy · effective September 14, 2026

What we keep, what we do with it, and how to make us let go of it.

Rebuild is made by NovaTrak. It turns the photos you already have into a contents inventory for an insurance claim, which means we handle pictures of the inside of your home. This policy says plainly what we collect, what we do with it, who else touches it, and how to have it deleted.

It covers the Rebuild web app, the pages around it, the links you share from it, and the emails it sends. Write to info@novatrak.com with any question about it.

1. Who we are

NovaTrak makes and operates Rebuild and is responsible for the personal information described here. You can reach us at info@novatrak.com.

2. What we collect

We collect only what the work needs. In order of how it arrives:

  • Your account. Sign-in is handled by Clerk, a sign-in service. Clerk gives us your name, email address and profile picture, and we keep a copy of those. Your password never reaches us.
  • Your photos. Your browser opens your photo export on your own device. For each photo it sends us three small copies: a 512-pixel version for the model, a 320-pixel thumbnail, and a 1280-pixel view for the screen. The full-resolution original is not uploaded and never leaves your device. We also store the file's name and its folder path inside the export (for example Takeout/Google Photos/IMG_0012.jpg), its original size, and the name of the archive it came from. Those can carry dates and names, so treat them as part of your data.
  • Photos from people you invite. When you send someone an upload link, their browser does the same: three small copies per photo, plus the name they type. Their photos are held for your approval and are not scanned or counted until you approve them.
  • What the scan finds. For each photo: whether it shows the inside of your home, the room, how sure the model was, and a one-line reason; and for each item seen, a description, its attributes, where it sits in the photo, and a confidence score. We keep your corrections beside the model's original values.
  • The inventory and the claim. The belongings built from those items, their estimated values, the facts you add (year bought, condition, what it cost), and the details you type for the claim document, such as your carrier, policy number, adjuster, key dates and the property address.
  • Receipts, documents and estimates. Receipts, warranties and purchase records you attach, and any carrier estimate you upload for comparison, are stored whole.
  • Payments. Paid plans are bought through Stripe. We store your Stripe customer id, the plan bought, the amounts, and any discount code used. Card numbers go to Stripe and never to us.
  • Usage. Each model call is recorded with what it was for and what it cost, so plan allowances and the spend safety brake can work. Background jobs and their progress are recorded too.
  • Technical data. Our hosting provider keeps ordinary server logs (IP address, browser type, time of request). On the live site we use Google Analytics to measure how the pages are used, and Meta's pixel on our public pages to measure our Facebook and Instagram ads.

3. How we use it

We use the information to:

  • Run the service: scan your photos, group what was seen into belongings, price them, build the claim document, and make the files you export.
  • Apply your plan: count photos against its cap, and pause model work when the spend safety brake trips.
  • Talk to you: email you when a long job finishes, send the invitations you ask us to send, and answer you when you write.
  • Take payment and keep the records that tax and accounting rules require.
  • Keep the service safe: spot abuse, enforce link budgets, and investigate problems.
  • Improve Rebuild, using aggregate figures that do not identify a person.

We do not sell personal information. We do not show advertising. We do not use your photos to train models, ours or anyone else's.

4. How the model is used

The scan, the grouping, the pricing, receipt reading and estimate matching are done by asking a model run by OpenAI. The 512-pixel copy of each photo is sent with an instruction to list the belongings in view. The model is asked about things, not people; it is never asked who is in a photo. Photos the model judges were not taken inside your home are set aside and used for nothing else.

Receipts and carrier estimate files are sent to the model whole so their lines can be read. Descriptions of your belongings are sent as text for grouping and pricing.

Under OpenAI's API terms, data sent through the API is not used to train OpenAI's models. What the model returns is an estimate and can be wrong. That is why every result in Rebuild can be corrected, left out, or deleted, and why the documents it produces are drafts for you to review, not appraisals.

5. Who else receives data

We use a small number of service providers. Each receives only what its job needs and acts under its own terms and security commitments:

  • Clerk, for sign-in and account security.
  • Supabase, for our database and for private file storage of the photo copies, receipts, estimates and exports.
  • OpenAI, for the model described above.
  • Stripe, for payments.
  • Resend, for the emails we send you.
  • Vercel, for hosting and running the app.
  • Google Analytics, for usage measurement on the live site.
  • Meta, to measure whether our Facebook and Instagram ads reach people. Its pixel runs on our public pages (the home page, pricing, the pages our ads link to and the like), never inside the app, on sign-in, or on a claim or upload link someone shares. If you visit one of those public pages and then create an account, we tell Meta that an account was started, with your email address and account id in hashed (one-way scrambled) form, your IP address, your browser type and the ad-click identifiers its pixel set. Meta never receives your photos, your inventory or anything else inside the app.

Beyond providers, data goes only where you send it: anyone holding a share link you create can read that claim document; people you send an upload link can add photos to your household; and the files you export are yours to hand to your carrier or anyone else. We may also disclose information when the law requires it, or to protect the rights and safety of you, us, or others. If the business changes hands, this data would go with it under this policy.

6. Cookies

Clerk sets the cookies that keep you signed in; the app does not work without them. On the live site Google Analytics sets cookies to measure use; you can block them in your browser or with Google's opt-out tools and Rebuild works the same. When you arrive from a link that names where it came from (an ad, a partner's page, a link a friend passed on) or open one of our pages for a particular community, we keep that source and any discount code the page carries in our own cookies for 90 days, so the code is ready at checkout and we can tell which partners reach people. The source goes to Stripe with a purchase. Blocking these cookies only means typing the code yourself. On our public pages, Meta's pixel sets advertising cookies (_fbp, and _fbc when you came from an ad) so Meta can tell whether an ad led to an account; the app itself never loads it, and you can block these cookies in your browser and Rebuild works the same. Fonts are served from our own servers, so no page calls a third-party font service.

7. Retention and deletion

We keep your data for as long as your account exists, because a claim can take months and you may need to come back to it.

Start over, in Settings, deletes the household's photos, scan results, inventory, claim details, receipts, estimates and saved files. Deleting your account deletes all of that and your profile. Once deleted, data leaves our provider's backups on their own schedule, within a limited period. We keep payment records for as long as tax and accounting rules require, and usage figures in aggregate form that does not identify you.

8. Security

Data moves over encrypted connections. Files sit in a private storage bucket that is not publicly readable; every row and file is keyed to your household and reached only through the signed-in session or a link you created. Staff access is limited to what running and supporting the service requires. No system is perfect; if you believe something has gone wrong, write to us and we will look at once.

9. Your choices and rights

You can act on your data yourself, at any time:

  • See it: the app shows everything it holds about your household, and the CSV and JSON exports give it to you as files.
  • Correct it: every description, value and detail can be edited, and every item the model got wrong can be left out.
  • Delete it: Start over in Settings, or delete your account.
  • Limit analytics and ad measurement: block the cookies as described above.

Depending on where you live, you may also have legal rights to access, correct, delete or receive a copy of your personal information, to object to or restrict some uses, and not to be treated differently for exercising them. Write to info@novatrak.com and we will respond within the time the law allows. We may need to confirm it is you asking.

10. Children

Rebuild is for adults handling a household's insurance claim and is not directed at children. We do not knowingly collect personal information from anyone under 13. If you believe a child has given us information, write to us and we will delete it.

11. Where data lives

Our providers store and process data in the United States. If you use Rebuild from elsewhere, your information is transferred to and handled there under this policy.

12. Changes to this policy

When this policy changes, the new version is posted here with a new effective date. For a change that matters to how your data is handled, we will tell you by email or with a notice in the app, and may ask you to agree to the new version before continuing.

13. Contact

Questions, requests and concerns about privacy go to info@novatrak.com. The Contact page says what to include so we can help quickly.

Questions about this document? Write to us.