Trust & data
How Throttle works, and how it pays for itself
Last updated · Questions: [email protected]
Every business that is free to use is paid for by somebody. This page says who, in plain English, before you have to go looking. It is the short version of our privacy policy — the long version is binding, this one is readable, and they say the same thing.
1. The short version
If you read nothing else on this page.
- Browsing the market costs nothing and does not require an account. No ads anywhere on a listing page, ever.
- We never sell your personal information. Not your email, not your name, not your saved searches, not your watchlist, not a record of what you looked at.
- We do publish and license aggregated, de-identified market data — statistics about machines and prices, computed across thousands of people, with nothing in them that points back at any one of you. Dealers, lenders, insurers and researchers pay for that. It is a real part of how this is funded, so we say so here rather than burying it.
- We keep a log of how the site is used — which filters get used, which categories get looked at — with your identity replaced by a salted hash and never the words you typed into the search box.
- Our crawler reads dealer websites, says who it is, obeys
robots.txt, and stops within a day of being asked.
2. What is free
The useful part, not a teaser of it.
A free tool that is only a demo of the paid one wastes everybody's time. The core of Throttle — finding a machine, comparing it, and seeing whether the price moved — works without paying and without an account:
- Search and filter every listing we track, across every category.
- Every listing page, with its full specs and a link straight to the dealer.
- Decode a VIN, a HIN or an N-number.
- The last 30 days of price history on any listing.
- The forums, vehicle stories, profiles, points and badges. All of it.
- An account adds 10 saved listings, 1 saved search and 5 decoder lookups a day, free.
We are not planning to move any of that behind a paywall. What Throttle+ sells is the archive — history and comparisons that only exist because we have been watching for months — not access to the machines themselves.
3. What Throttle+ costs, and why
And why we are not currently selling it.
Throttle+ is $15 a month or $79 a year. It unlocks sold records and what they actually went for, the complete price history rather than the last 30 days, days-on-market, saved-search alerts, and the market analytics.
It is not on sale right now, on purpose. Sold history and price trends need roughly two months of daily crawling behind them before they mean anything. Until then those pages are structurally empty, and charging a subscription for an empty chart is the kind of thing that is hard to come back from. When the history is there, the buttons turn on. Not before.
There is also a one-time BETA Supporter offer — $250, capped at 500 people, closing Dec 31, 2026 or when the cap is reached, whichever comes first. It is deliberately framed as funding the build rather than pre-buying a subscription, because those are different promises and only one of them is honest today. Every limit, exclusion and the refund window is printed on /pricing before you pay, and again in the terms. The old “lifetime” plan is retired: an unbounded promise is the one that eventually has to be broken.
4. How we make money
One paragraph. No euphemisms.
Four things, and these are all of them. Throttle+ subscriptions at $15 a month to $79 a year, which is the main one and the one the product is built around; a limited one-time BETA Supporter offer of $250, capped and dated, which funds the first year of building; licensing aggregated, de-identified market data to businesses — dealers, lenders, insurers, manufacturers and researchers who want to know what a segment did last quarter, and who pay for statistics, not for people; and affiliate links on third-party tools, always labelled “Paid link”, which never influence ranking and are not live yet — when the first one ships, it ships already labelled and this sentence is already true.
We do not run ads — not on the site, not in the weekly email, not in the monthly market report. We do not take referral fees or commissions from dealers for sending you to them, which is why no dealer can pay to rank higher in a search result; an affiliate link to a tool is a different thing from a dealer paying for placement, and we will not blur them. And we do not sell your personal information, which is a distinct thing from the third item and is explained next.
5. What we record
Enough to make the product better and the statistics real.
When you use Throttle we write rows to a behaviour log: a search happened, a listing was viewed, something was saved, an alert fired, a forum post went up. We use it to fix what is broken, decide what to build, train our own internal models, and produce the market statistics described below.
Your identity is not in those rows. It is replaced by a salted hash — your user or session id run through a one-way function with a secret key that lives only on our server and is never exported with the data. Somebody who stole an entire export and a copy of our user table still could not rejoin the two. That is what makes these rows de-identified rather than merely re-labelled.
One deliberate exception: for signed-in visitors we keep the user id on the row as well. It is the only way to honour a deletion request, and a store we cannot erase from is a worse privacy position than one we can. Deleting your account deletes those rows.
6. What we never record
A hard list, enforced by the code that writes the log, not by good intentions.
| Never stored | What we keep instead |
|---|---|
| Your IP address | In the behaviour log, a two-letter country code and nothing else derived from it — the log never touches an address. Two places outside the log do need to tell requests apart: rate limiting and the audit trail. Those keep a salted one-way hash, never the address, and the salt is a secret that lives only on our server and is never exported |
| Your browser's user agent string | A coarse family such as safari-mobile or bot |
| Your email, name, phone number or anything you typed about yourself | Nothing. These never reach the log at all |
| The full page that referred you | The hostname only |
| The words you type in the search box | The shape of the search: which filters, how many results, how it was sorted |
The search box is the important one. It is where people type their own name, their VIN, the town they live in and occasionally their address, so it is the one field the log refuses outright. Event properties are flat values only, capped in number, and every key is screened for fragments like email, phone, address and vin — a key that trips the screen throws the whole event away rather than quietly cleaning it.
7. What “aggregated” actually means
The word does a lot of work, so here is what it is doing.
The data products we license look like this: the median advertised price of a model year in a region and how it moved over a quarter; typical days-on-market by category; how often listings in a segment get discounted and by how much; demand expressed as an indexed, rounded count across thousands of visitors.
They do not look like this: a list of people, a list of what one person searched for, an email address, or any row a buyer could match back to you. Where a statistic would be computed from too few people to hide anyone in, the cell is suppressed or widened until it is not.
We hold ourselves to the CCPA/CPRA standard for de-identified data, which is four things at once:
- the data cannot reasonably be linked back to a person;
- we publicly commit never to attempt to re-identify it — this sentence is that commitment;
- we keep the technical controls that make it true: the secret salt, the forbidden-field list above, no raw addresses, no search text;
- every business we license it to is contractually bound to the same no-re-identification rule in the Data Licence, and loses the licence if it breaks it.
If we ever wanted to do something that amounted to selling personal information, we would have to change the Privacy Policy first, tell you, and offer an opt-out before it started.
8. The promises
Four, each one checkable.
No ads on listing pages
Not a banner, not a sponsored unit, not a “featured” slot a dealer paid for. The order of search results is decided by your filters and nothing else. A dealer cannot buy a better position, and there is no contact form between you and the store — we link you straight to their page.
We never sell your personal information
In the sense the CCPA and CPRA define it, and we do not share it for cross-context behavioural advertising either. Your email, your handle, your saved searches, your watchlist and your event rows do not go to a data broker, an advertiser or a dealer, for money or for anything else of value.
An honest crawler
ThrottleBot identifies itself in every request, reads robots.txt before anything else and obeys it, makes one request at a time per site with at least 2.5 seconds between them, and stops when it sees a block rather than changing its name to get around one. We index dealers' own websites — never aggregators, never anything behind a login. The details are on the ThrottleBot page.
Dealers can opt out, in one email
Any dealer can email [email protected] from an address at their own domain and be removed within 24 hours, with written confirmation. No form, no retention offer, no argument. A robots.txt block works on its own without contacting anyone.
9. What we are not promising
The part most pages like this leave out.
- That the prices are right. They come from dealer websites and can be stale, mistyped or wrong at the source. Always confirm with the seller before you drive anywhere.
- That the prices we extracted are right. As of 21 September 2026 we know some are not — our own parsers, not the dealers, produced wrong prices and false discounts in the current index. One price was copied across a dealer’s whole inventory and discounts were invented against it. Crawling is paused and we are re-crawling from scratch rather than patching the rows, because a price history built on fabricated prices is not salvageable. Market statistics are switched off until then rather than computed from it. This bullet is dated and will be removed when it stops being true, not quietly softened.
- That we will exist forever. We are small and self-funded. Supporter access lasts as long as the service does, which is the honest formulation and the reason the offer is capped and refundable rather than sold as a perpetuity.
- That the free tier will never change. We have no plan to gut it and every incentive not to. But we would rather say that than promise “forever” and have to explain ourselves later.
- That points and badges are worth money. They are not. They have no cash value, cannot be sold or transferred, and we may change how they work.
10. What you control
All of it is self-serve.
- Global Privacy Control and Do Not Track. If your browser sends either signal, we honour it — the behaviour log stops recording you. We do not make you find a setting to do what your browser already said.
- Delete your account. From your settings. It removes your profile, saved searches, watchlist and the user id on your event rows.
- Ask what we hold. Email us and we will tell you, and export it.
- Your full rights — access, correction, deletion, portability, and the rest — are in the privacy policy, with the addresses to use.
11. Tell us we got it wrong
If something on this page does not match what Throttle actually does, that is the most useful bug report we can get, and we would rather hear it from you than read it somewhere else. Email [email protected].
The binding documents are the Terms, the Privacy Policy, the Data Licence, the API Terms and the Community Guidelines.