Privacy Policy
What we collect, why we collect it, and the choices you have. In short: little, for clear reasons, and never sold.
Effective date: 21 November 2026
Summary in plain language
- We collect the least we need to run a waitlist, let people sign in and give credit for verified work.
- Your AI provider keys stay on your machine. They are never sent to us.
- You appear under a username (a handle). Your real name is shown only if you choose that.
- We do not sell your personal information and we do not share it for cross-context advertising. There are no ad trackers on this site.
- We use essential cookies only, and our analytics is cookieless.
- You can ask us to give you a copy of your data, correct it or delete it, by writing to hello@sparetokens.ai.
SpareTokens is a private pilot. This policy describes what the site, the waitlist and the worker actually do today, including sign-in. It sits alongside our Terms of Service. SpareTokens is run by Paul Ferguson, 638 Lindero Canyon Rd #422, Oak Park, CA 91377, USA, and is the controller of the personal data described here.
What we collect
Waitlist and updates
When you use the form on /join we store your email address, the optional boxes you tick (the providers you could contribute and whether you want to vote on future challenges), the time you signed up, whether you confirmed your email, your personal invite code and an optional referral tag. The form uses double opt-in: we email you a confirmation link before you count, and every email carries an unsubscribe link that works without signing in. These records are kept in Cloudflare KV.
Abuse prevention
To rate limit the form we use a hashed form of your IP address, not the address itself. To cap how many emails any one mailbox receives we also keep a one-way hash of the address with a counter for about an hour. Cloudflare Turnstile checks that a visitor is a person, on the join page, the sign-in page and the waitlist form. Turnstile runs in your browser and may look at signals such as your browser and network to decide that. The coordinator's rate limits hold addresses in memory only, and its log files are rotated and deleted on a schedule (see Retention). We never store an IP address in a database or show one publicly.
Accounts and sign-in
You can sign in with GitHub, Google, a passkey or an email link. When you sign in or register we store a username (handle) and the identifier your sign-in method gives us: your GitHub user id and username, your Google account id, or your passkey's public key (the private key stays on your device). We store an email address only once it is verified, by GitHub (your primary verified address), by Google, or by you opening a link we email, and we record when it was verified. We never receive your password, and accounts are never joined because they share an email address. A session cookie keeps you signed in. Sessions are stored by a one-way hash of the cookie, last up to 30 days, and end when you sign out.
Profile, results and the leaderboard
You are shown under your username by default. Showing a real name, a link, a country, the models or providers you used, a hardware label, or your tokens and spend are each opt-in, with a visibility setting for each field. Verified results, task counts and your score are public under the name you chose, because they are the ranking. Email addresses and IP addresses are never public. Anonymous mode shows you as a numbered contributor.
The worker and task data
The worker runs on your own machine with your own credentials. Your provider API keys and logins are never sent to us. The worker sends the coordinator the results of tasks, the files a task produces (its artifacts, which we treat as task logs), and usage numbers it reports about itself, such as tokens and spend. Your full local transcripts of each task stay on your machine and are never uploaded. Those usage numbers are self-reported, are labelled that way wherever shown, and never affect rank. Results the worker returns are scanned locally for key-shaped strings before they leave your machine, but you should not put secrets into task output.
Phone companion
If you use the phone app at /app, the coordinator stores the small status summary each of your workers reports (running, paused or idle, and why), and your remote pause or boost requests. If you turn on notifications, we store your browser's push subscription and your notification choices. Turning notifications off removes the subscription, and deleting your account removes all of this.
Messages to us
If you write to hello@sparetokens.ai or support@sparetokens.ai we keep the message (its text, the sender address and the sender's email authentication results; attachments are not stored) and our reply so we can help you, for 90 days, and then delete them. A software assistant reads incoming messages to sort them and to draft answers to routine questions from our public documentation; it can only reply to the sender, unsubscribe or start a delete request for the sender's own address, or file an anonymous bug report. It cannot see anything else we hold, and anything that needs a person is passed to us. Obvious spam is dropped without being stored. Mail to hello@sparetokens.ai is also forwarded to the project owner's own mailbox. If an email we send to you bounces or you report it as spam, we keep that address on a do-not-email list so we do not send to it again.
Why we use it
- To run the waitlist, confirm your email and invite you to the beta.
- To send SpareTokens updates you asked for, and to honor unsubscribe requests.
- To provide accounts, keep you signed in and show your profile as you set it.
- To hand out tasks, verify results and credit verified work on the leaderboard.
- To prevent abuse, spam and fraud, and to keep the service secure.
- To understand, without cookies, which pages are read, so we can improve the site.
- To answer your questions and meet legal obligations.
We do not use your data to train AI models, and we do not make decisions about you by automated means that have legal or similarly significant effects.
Legal bases (GDPR)
If the GDPR or the UK GDPR applies to you, we rely on these bases.
- Consent: the waitlist and updates emails, and any optional profile details you choose to show. You can withdraw consent at any time, for example by unsubscribing or changing your profile.
- Contract: providing your account, running tasks and crediting verified results, which is what you ask for when you contribute.
- Legitimate interests: securing the service, preventing abuse, rate limiting, cookieless analytics and keeping the leaderboard accurate, balanced against your rights.
- Legal obligation: where the law requires us to keep or disclose information.
California notice (CCPA and CPRA)
In the last 12 months we have collected these categories of personal information: identifiers (email address, username, hashed IP address, invite code), internet activity (pages viewed in cookieless analytics, usage numbers the worker reports) and, if you opt in, a name, link and country. We collect them from you and from the sign-in provider you pick, and we use them for the purposes in this policy.
We do not sell personal information and we do not share it for cross-context behavioral advertising, so there is nothing to opt out of. We do not knowingly sell or share the personal information of people under 16. We do not use or disclose sensitive personal information to infer characteristics about you. California residents may ask us to tell them what we hold, to delete it, to correct it, and may do so without being treated differently. We will verify your request by confirming control of the email address on the account. You can also use an authorized agent.
Processors and who sees your data
We use a small number of service providers who process data for us under their own terms and privacy policies.
- Cloudflare: hosting, the network edge, KV storage for waitlist and account records, Turnstile bot checks, cookieless Web Analytics, and Email Routing, which receives mail sent to our addresses and hands it to us.
- SendGrid (Twilio): sending confirmation and update emails. Messages are sent as they are requested, and each one carries an unsubscribe link.
- Anthropic: processes the text of messages sent to support@ so that the assistant described under "Messages to us" can sort them and draft answers. It does not receive our other records.
- GitHub and Google: only if you choose to sign in with them. They tell us who you are and we tell them you signed in.
The coordinator runs on our own infrastructure. We do not sell data, and we disclose it to others only when the law requires it, to protect the service or its users, or in a business transfer with the same protections. Public profile fields are visible to anyone.
Cookies and similar technologies
We use essential cookies only, such as the session cookie that keeps you signed in, and short-lived cookies that protect a sign-in while it is in progress (including one that ties an emailed sign-in link to the browser that asked for it). We do not use advertising or tracking cookies. Cloudflare Web Analytics is cookieless: it counts page views and referrers in aggregate without setting cookies, without personal data and without tracking you across sites. It is not loaded on the sign-in, dashboard, phone app, invite, unsubscribe or data-deletion pages. Cloudflare Turnstile runs on the join and sign-in pages to stop bots. Because we use essential cookies only, we do not show a cookie banner. Our site also works with JavaScript off for reading.
Retention
- Task logs: 30 days. This covers the files a worker uploads with a result that did not verify, which are cleared after 30 days, and the full transcripts kept on your own machine, which the worker deletes after 30 days. Verified results and their evidence are part of the public record and are kept.
- IP addresses: 7 days. The signup form's rate limit stores a hashed address for about an hour. The coordinator's rate limits keep addresses in memory only and drop them after 7 days without use. Its log files are rotated daily and rotated copies are deleted after 7 days. Our host, Cloudflare, keeps its own request logs under its own terms.
- Account data: until you delete your account.
- Messages to hello@ and support@: 90 days, then deleted automatically. The do-not-email list (addresses that bounced, were reported as spam or unsubscribed) is kept so we do not email you again.
- Waitlist records: until you unsubscribe and ask us to delete them, or we close the waitlist. After you unsubscribe we keep a minimal record so we do not email you again.
- Verified results: kept as part of the public record. When you delete your account, they stay and are shown as "Former contributor", with no name, link or country.
Your rights and how to use them
You can ask us for access to your data, a copy you can take with you (export), correction of anything wrong, and deletion. Depending on where you live you may also have the right to object to or limit some processing, to withdraw consent, and to complain to your data protection authority. You can do most of this yourself. Signed in, your dashboard lets you download everything we hold about your account as a JSON file (access and export), edit your profile and sign-in methods (correction), and delete your account (deletion), which needs a sign-in within the last hour and a typed confirmation. If you only joined the waitlist or the updates list, use delete my data: we email a link to your address and pressing the button erases your records. The link usually arrives within a minute. If it does not, write to hello@sparetokens.ai from your address instead. Every email also has an unsubscribe link that works without signing in. We reply to written requests within 30 days, and sooner when we can. Deleting your account removes your name, link, country, sign-in methods, sessions, worker tokens, reports you wrote and your email records, and unsubscribes your email. Verified results stay attributed to "Former contributor".
Children
SpareTokens is not directed at anyone under 16, and you must be at least 18 to contribute because you use your own AI accounts and pay for your own usage. We do not knowingly collect personal information from people under 16. If you think a child has given us data, write to us and we will delete it.
International transfers
We are in the United States and our providers process data in the United States and other countries. If you are in the European Economic Area, the United Kingdom or Switzerland, your data is transferred to and stored outside your country. Where the law requires, we rely on safeguards such as the standard contractual clauses and our providers' data processing terms.
Security
We limit what we collect, keep provider keys off our systems by design, keep IP addresses used for rate limiting only as hashes or in memory, sign confirmation and unsubscribe links, and serve the site over HTTPS. No system is perfectly secure. If a breach affects your data we will tell you as the law requires. You can read our approach on the security and trust page.
Changes to this policy
We will update this page when what we do changes and revise the effective date above. If a change is material we will say so on the site or by email to people on our list, before it takes effect where we can.
Contact
Email hello@sparetokens.ai, or write to SpareTokens, 638 Lindero Canyon Rd #422, Oak Park, CA 91377, USA. See also the Terms of Service and the accounts and privacy docs.