What this site collects,and what it refuses to

This page explains what Read the Nameplate collects when you use the Quote X-Ray or the rest of the site, why, and where it goes. It's written to match what the code actually does — not a generic template — so if something here ever stops being true, that's a bug, not fine print.

What happens when you paste or upload a proposal

If you upload a PDF, the file itself never leaves your device. Text is pulled out of it in your browser; only that extracted text — not the file — is used from that point on. If it's a scan or a photo with no real text layer, nothing is extracted and nothing is sent anywhere.

When you submit the text (pasted or extracted), it's sent to our server over an encrypted connection. The first thing that happens to it there, before anything else touches it, is redaction: emails, phone numbers, street addresses, full ZIP codes, and account-style numbers are stripped automatically. Names are handled on a best-effort basis — a name sitting in a labeled line, a salutation, or a signature block is caught reliably; a name mentioned elsewhere in the body of the document may not be. Don't paste anything containing information you wouldn't want to see reach an AI model, even redacted.

Who reads the redacted text

The redacted text is sent to Anthropic's Claude API to generate the reading — what the document says, what it's missing, and the plain-language explanation. Anthropic processes this as our service provider; see Anthropic's own privacy policy for how they handle data sent through their API. Every number shown to you (price position, percentile, comparison) is computed by ordinary arithmetic on our own server, never by the model — the model only reads and writes, it never calculates a statistic.

What gets stored, and what never does

Storing anything is optional, and you control it with a single box on the quote form. "Add this quote to the comparison dataset" is checked by default — the comparison this site runs on is built entirely out of quotes other readers contributed, so contributing is the normal case. Uncheck it and nothing about your submission is saved anywhere once the response is sent back to you. What that dataset is then used for — including the combined statistics this site sells to pay for itself — is set out under Selling and licensing the aggregate data below.

If you leave it checked, what's stored is:

• Job type (furnace, AC, heat pump, etc.)
• Equipment description and capacity, redacted the same way as above
• Efficiency rating, if stated
• Total price
• Which scope items the document included or omitted
• State, and only the first three digits of the ZIP code (roughly a metro area, not an address)
• The date on the proposal, if stated

What's never stored, checked or not: the proposal text itself, the original PDF, your name, your address, your phone number or email, or the contractor's name or business. Only the structured fields above ever reach the database, and only while that box is left checked.

IP addresses

Requests to the Quote X-Ray are rate-limited to prevent abuse. That limiter checks your IP address against a short-term counter held in server memory — it's never written to a database, logged permanently, or associated with anything you submit.

Cookies and analytics

None are currently in use on this site. If that changes, this page will say so, with what's tracked and why.

The email newsletter

Subscribing to new-post emails is entirely separate from the quote tool and entirely optional. It's handled by Buttondown, who has its own privacy policy. Your email address is used only to send you new posts; unsubscribe any time from the link in every email.

Who else sees data, and why

The service providers this site relies on to run at all:

• Anthropic — reads the redacted proposal text to generate the reading (see above)
• Neon — hosts the database that stores the structured, contributed rows described above
• Vercel — hosts the site itself
• Buttondown — sends the optional email newsletter, if you subscribe

None of these receive your name, address, phone number, or email through the quote tool — only through the separate, optional newsletter signup, and only Buttondown sees that.

Selling and licensing the aggregate data

This site is free and carries no ads. It's paid for by selling market intelligence built from the combined dataset — medians, price-per-ton figures, regional averages, brand-versus-brand installed cost, and trends over time — to companies in the industry: efficiency program administrators, equipment manufacturers, insurers, lenders, and researchers. You're being told this here rather than discovering it later, because a site whose whole premise is that documents should be readable doesn't get to bury its own business model.

What's sold is always combined statistics, never your individual row. Nobody buys a list of quotes. And because a submission carries no name, address, email, phone number, contractor, or proposal text — and no geography finer than your state and the first three digits of your ZIP code — there is nothing in the product that identifies you or your home. Anyone licensing it is barred by contract from attempting to re-identify a submission.

We don't sell leads or referrals to contractors, and we won't — that's a standing promise on the About page, not a policy that changes when it gets inconvenient. The full grant you give by leaving the contribute box checked, and the list of things it never permits, is set out in the Terms of Service.

How long contributed data is kept

Structured rows added to the comparison dataset are kept indefinitely — that's what makes the comparison useful to the next person. Because submissions aren't tied to an account, an email, or anything else identifying, there's generally no way to look up which row was yours after the fact. If you have a specific concern about a submission, contact us (below) and we'll do what we reasonably can.

Children

This site isn't directed at children, and we don't knowingly collect anything from anyone under 13.

Changes to this policy

If what the site actually does changes, this page will be updated to match, with the date at the top revised. Meaningful changes will be noted in the blog.

Contact

Questions about this policy, or a request related to something you submitted: privacy@readthenameplate.com.

Read the Nameplate is published by Jaxter Labs LLC. See also the Terms of Service and the scope disclaimer at the bottom of every page.