Most AI tools log everything — your questions, your IP, your habits. ZT-AI is built to keep no record of what you do here. Below is exactly what that means, in plain language — including the limits, stated honestly. If any line here ever stops being true, that's a bug, and we want to hear about it.
Last updated: July 2, 2026
We never ask for your name, email, or any personal information to use ZT-AI. There is no sign-up form, no login page, no username, no password, no profile. Free use is completely anonymous — you're just a request, and then you're gone.
Your questions are sent to our server, routed to the AI model, and the answer is returned to you. That's it. We do not store what you asked. We do not build a history. We do not analyze your questions over time. Once your result comes back, the query is gone from our system — it only ever existed in memory for the moment it took to answer. When you close the tab, the conversation is gone from your side too.
To keep free use anonymous while still enforcing a daily limit, we count requests using a temporary tag held only in our server's memory. The tag is a hash of your IP combined with a random key that is regenerated every day and never written to disk or a database. Your raw IP is used for the instant it takes to build that tag and is then discarded — we never save it anywhere. When the day rolls over, the key is thrown away, so the previous day's tags can't be traced back to anyone, including by us. We don't use cookies to track you across sessions. No fingerprinting. No ad targeting.
We use GoatCounter for basic page-view analytics — a privacy-respecting tool that doesn't use cookies, doesn't store IP addresses, and has no way to identify individual visitors. We only see aggregate numbers like "200 people visited the pricing page," never who they are. This is the one analytics script that loads in your browser, and it's disclosed here on purpose.
When you pay, your card is handled entirely by Stripe. We never see or store your card details. But Stripe necessarily learns your billing identity — your name, email, and card — exactly like any online purchase. That's unavoidable with card payments.
For a subscription, we keep the minimum link needed to renew and cancel your plan. That link lives inside your Stripe payment record (Stripe already knows who you are), not in a table on our servers. The honest consequence: a subscription can, in principle, be traced back through Stripe to the person paying for it.
Here's the part that matters. That link can reveal only that you pay for ZT-AI. It reveals nothing about what you use it for — because we never record your questions or answers, and our AI providers do not retain them under their API terms. Membership is knowable. Activity is not.
Want the most anonymous paid option? A one-time query pack stores no subscription link on our side.
Your access token is a randomly generated string like zt_a3f9b2c1.... On our side we store just the token, its tier (Free / Pro / Ultra), the date it was created, its expiry, and any bonus balances you've bought (AI-only and live-web counts) — plus, for subscribers only, a one-way hash of the Stripe subscription ID used to end the plan on cancellation. For free and query-pack users there is no link to your identity on our side at all. For subscribers, the link back to your identity lives at Stripe, not in our database.
Keep your token safe — do not share it. If you believe your token has been compromised, contact us and we will issue a new one.
We don't run Google Analytics, the Meta Pixel, or any advertising or cross-site tracking scripts. The only third-party scripts your browser loads are Tailwind (for page styling), marked (for formatting AI answers on the chat page), and GoatCounter (the cookieless analytics described above) — none of them ad networks or behavioral trackers. Like any site that loads scripts from a CDN, those providers see the IP address your browser connects from (that's how the web works), but we send no data about you to any advertising network.
Answers are generated server-side by two AI providers, and we name them on purpose. AI-only queries go to Google's Gemini API on its paid tier, where under Google's terms your prompts and responses are not used to train Google's models; Google states it may retain API logs briefly for abuse and policy monitoring, then delete them. Live web searches go to the Perplexity Sonar API, which runs under Zero Data Retention: prompts and answers are deleted immediately after processing and are not used to train models. Your requests reach both providers from our server with no account or identity of yours attached.
In fairness: those are Google's and Perplexity's policies, not laws of physics. Either company could change its agreement in the future, and that is outside our control. We can promise what we do with your data; we can't make promises on another company's behalf — so we tell you exactly who touches it and under what terms.
Your browser stores two small items on your own device using localStorage:
zt_access_token — your token, saved so you don't have to re-enter it every visit. When you send a request, this token is sent to our server to check which plan you're on — that's its whole purpose. It's the same token you already hold, not new information about you. Clear it anytime with the "Clear" button in the token panel.zt_terms_accepted — just the value "true" so the Terms popup doesn't appear every visit. This one never leaves your device. No timestamp, no personal data.That's it. Two keys. You can verify this yourself: browser DevTools → Application → Local Storage.
⚠️ Incognito mode: If you use private/incognito browsing, both values are cleared when you close the window. You'll need to re-enter your token each session — but you get maximum privacy.
Like every website on earth, when you connect to zerotraceai.org your internet provider can see that you connected to us, and our web host briefly handles that connection to serve you the page. That's how the internet works. What we control is what we store — and we store no logs of your activity or conversations. Our hosting provider may keep basic operational logs (things like request timing and error traces) as any host does; these never include your chat content and aren't used to identify you. If you want to hide even the fact that you visited, use a VPN or Tor. That's a layer no website can provide for you, and we'd rather say so than imply otherwise.
In the spirit of full transparency, here is everything kept in our database — it's a single table of access tokens:
What we explicitly do not store:
That's the complete list. No search history. No behavioral data. No profiles.
Legal
All content, code, design, branding, and materials on this site are the property of Zero Trace AI and are protected by applicable copyright law. Unauthorized reproduction, distribution, or use of any part of this site is strictly prohibited.
You may not reverse engineer, scrape, crawl, or attempt to extract our system prompts, source code, or architecture. You may not attempt to bypass query limits, spoof tokens, or abuse our API in any way.
Violations may result in immediate termination of access and legal action. We take unauthorized use of our platform seriously and will pursue all remedies available under applicable law.
This Privacy Policy and any disputes arising from it shall be governed by the laws of the State of Oregon, United States, without regard to conflict of law provisions. Any legal action shall be brought exclusively in the courts of Oregon.
We may update this Privacy Policy from time to time. Changes will be posted on this page with an updated date. Continued use of the service after changes constitutes acceptance of the updated policy.
Questions about this Privacy Policy? Contact us here.
Ready to search privately?
Zero chat logs. Zero History. Zero Ad Tracking. Zero Trace of your questions.