Legal
Privacy policy
This page sets out the personal data ReserveBeat holds, why, and for how long. In short: an email address if you subscribe or open an API account, product analytics with automatic capture off, counters named after your network address, kept for at most a day, and the web server's access log, deleted 30 days after each file fills. Nothing else.
Who holds this data
The site is published by ReserveBeat LLC, which decides what is collected and why. Questions about this policy, and requests about your own data, go to contact@reservebeat.com.
Postal address: PO Box 620672, Charlotte, NC 28262.
What is collected
This list was written from the code that does the collecting rather than from a template. If something is not on it, this site does not collect it.
Alert subscriptions: your email address, which token you asked about, and the times you subscribed, confirmed and unsubscribed. Nothing is sent to an address that has not confirmed, apart from that one confirmation request.
API accounts: your email address, a one-way hash of your password, and the date the account was opened. Every API key is stored as a hash beside a short visible prefix; the key itself is never kept, which is why a lost key can be replaced but not recovered.
API usage: for each authenticated request, the time, the account, the key, the route, the response status and the compute units it cost. These are the records the monthly free allowance and a prepaid balance are counted from.
An audit record of account and subscription events. Since 27 July 2026 those records carry a one-way hash of the email address rather than the address itself.
Product analytics: named events this site chooses to send, and never automatic capture of clicks, and no automatic page-view capture by the analytics library — both are switched off; page views are the named event described below. Events travel through this site’s own domain, and your cookies for this site are stripped before anything reaches the analytics provider. Two further named events are sent: how fast each page loaded in your browser (the Core Web Vitals figures, with the page's route group rather than its address), and, once per visit, which site or AI assistant linked you here, read from the referring address or the campaign tag on the link; neither carries your address or your name. Requests to the site's MCP endpoint are counted by the kind of request, the tool asked for and the name of the client program, mapped to a short fixed list; what was asked for is never sent. Page views are counted as named events carrying the page path only — never your address or your name.
Sign-in attempts, sign-ups and API reads: the network address you connect from. Signing in, opening an account, asking for a password reset, changing a password or email address, closing an account, and signing in to the operator portal count attempts per address so that repeated wrong passwords can be slowed down, and that counter is discarded about twelve minutes after your last attempt. Once each is switched on, signing in with Google and asking for a code to connect an AI agent are counted the same way. Opening an API account is also counted per address per day (UTC) so that one network cannot open accounts without limit, and that counter is discarded within a day of the first attempt it counts. Reads of the API under /v1 are counted per address so that one caller cannot take the whole service, and that counter is discarded about two minutes after your last read. In each case the address is held only as the name of a counter in the short-term store that does the counting, and the address is never written to the database, never written to a log line, and never joined to an account or a subscription. Nothing else in the application records a network address. The web server in front of this site keeps its own access log: one line per request with the time, your network address, the request's method, host, path and protocol, every request header your browser sends (your browser's identifying string and the page that linked you here among them; any query string is removed from the path, the referring page and any redirect, and a sign-in credential or cookie is written as redacted), the response's status, size and headers, and how long it took. It exists to count which pages, crawlers and AI bots reach the site and to trace faults. Each log file fills at 20 MiB, at most five filled files are kept, and a filled file is deleted 30 days after it fills, so a line is held until its file fills and for 30 days after that, and the whole log never exceeds about 120 MiB. It is not shared with anyone, with one exception: a request whose browser identifying string names a known crawler, AI bot or automated client is counted in the analytics service below as one event carrying that agent's name and kind, the page path without its query, and the response status — never the network address, and never anything from a request made by an ordinary browser. It is never joined to any account or subscription.
Why it is collected
Addresses exist so that a confirmation or an alert can be delivered and an account can be signed in to. Usage records exist so that an allowance and a balance can be counted and a key can be revoked when it is misused. Analytics exist so that we can tell which pages are read. None of it is sold, and none of it is used for advertising.
Session recording
Session recording starts switched off and turns on only while you are reading the pricing page, and only if you have asked this browser to be remembered. That is the whole list of paths it covers; it is off everywhere else on this site, including every page that carries a figure.
Cookies and local storage
Signing in to an API account sets one cookie holding your session token. It is readable only by the server, expires twelve hours after it is issued, and is deleted when you sign out. Signing in to the operator portal sets an equivalent session cookie. Signing in with Google also sets, for at most ten minutes, a cookie that ties Google's answer to the browser that asked for it; it is readable only by the server and is deleted when Google sends you back.
Analytics runs without a cookie unless you ask for one. By default the anonymous identifier the analytics library uses is held in memory for as long as the page is open and is never written to disk, so no analytics cookie is set and every visit is counted as a new one. The home page asks once, in a single line at the foot of the page, whether to remember this browser; declining changes nothing and the question is not asked again. If you ask this browser to be remembered — there, or with the setting below — the library then sets its own cookie under this site’s domain to hold that identifier between visits. With no analytics key configured, no analytics cookie is ever set, nothing is sent, and the question is never asked.
The one-question survey records in your browser’s local storage the date it was last shown, so that it is not shown again within a week. Your analytics setting is recorded in the same place, and only once you change it from the default. The home page’s one-time question records there that it has been answered, so that it is asked once and not again, and your colour-theme choice is recorded there too. None of these values ever leaves your browser.
How long it is kept
Unsubscribing sets the subscription to unsubscribed and stamps the time. The row and the address stay, because that record is the proof the request was honoured. Nothing further is sent to an unsubscribed address unless you subscribe again.
Audit records and API usage records are append-only by design: they cannot be edited or deleted, by anyone, including by us. That is what lets them stand as evidence, and it is also a real limit on erasure rather than a form of words.
No audit record on this service holds an email address in plain text; a check of every audit record on 10 October 2026 found none. Audit records cannot be removed, so the one-way hash of an address stays in them after the account it belongs to is closed or deleted. We state that here rather than promise an erasure that this system cannot perform.
Apart from unconfirmed API accounts, no scheduled job deletes account, subscription, key or usage records. The retention sweep that does run purges evidence artefacts, which hold no personal data. An API account whose email address is never confirmed is deleted 7 days after it was opened, with its confirmation link. Audit records about it are append-only and remain; they carry a one-way hash of the address, not the address.
Who else can see it
Each service below is reached only when it has been configured. Where one is not configured, nothing reaches it at all and the feature it powers is simply absent.
PostHog, in the United States: the named product events this site chooses to send, and session recordings from the one page they are switched on for, only from a browser that has asked to be remembered; and a count of requests from known crawlers, AI bots and automated clients, by agent name and page path, with no network address. Cookies for this site are stripped before any request reaches it.
An email relay, for confirmations and alerts. Your address travels in the message envelope and nowhere else — it is deliberately kept out of error text, queue records and logs.
Stripe, once purchases of compute units are switched on: your account email address, the compute units you bought, and any campaign parameters carried by the link you arrived through. Card details go to Stripe and never to us. From Stripe we keep, on the record of each purchase, Stripe’s reference for the payment and a one-way hash of the identifier Stripe gives the card that paid — never the card number — so that each account is funded by its own card.
Google, once signing in with Google is switched on, and only when you choose it: Google tells us the email address on your Google account, that Google has verified it, and an identifier for your Google account. We keep the address as your account's address and the identifier on your account record, so that the same Google account signs in to the same API account; closing the account removes both. We ask Google for nothing else — no name, no picture, no contacts — and we keep none of the tokens Google sends. Google sees that you are signing in to this site, under Google's own privacy policy.
Formbricks, once the one-question survey is switched on: the answer you give it.
DigitalOcean, which hosts the servers and therefore holds at rest everything this site stores.
The list is limited to outside services that can see a visitor's or an account holder's data. Services that no such data crosses are not processors of it and are not listed: the content and link tooling (OpenRouter, Dub, IndexNow), the operator's own alerting, and every public source the service reads from — blockchains, the issuers' own published reports, timestamping services, regulators and public data aggregators.
Your choices
You can unsubscribe from an alert using the link in its footer, close an API account, or ask what is held about you. Requests about your own data go to contact@reservebeat.com.
A signed-in API account can also be closed from the dashboard, without writing to us. Closing revokes every API key and webhook endpoint on the account at once and removes the email address from the account record, and a closed account is never reopened; the same address can open a new account later. Audit and usage records are append-only and survive the close, and so do the revoked key and endpoint records they refer to. Audit records written since 27 July 2026 carry the address as a one-way hash, never the address itself.
Two limits are stated plainly rather than buried. Append-only records cannot be erased, and an unsubscribed row is kept as the record of the request. Everything else that can be removed on request will be.
The corrections path, and what happens after a message arrives, are set out on the contact page. Contact
Changes to this policy
This policy changes by supersession, like the terms of service: a new dated version replaces the previous one, and the previous one is kept on the record rather than deleted.