Effective date: April 9, 2026 · Last updated: June 14, 2026
Get Real is a conversation card game. It doesn't need your name, your email, or an account — and it doesn't ask for them. This policy describes exactly what information is collected, what isn't, and how the AI chat feature works.
Short version: We don't collect personal information. The app stores only your settings on your own device. If you use the AI chat feature, the text you type is sent to Anthropic to generate a response. We never store the content of your chat messages. We do keep a small operational log of how many tokens each chat call used, so we can monitor cost and prevent abuse — details below.
§ Contents
No account required
Get Real requires no registration, login, or account of any kind. You are not identified, tracked, or profiled. You can use every feature of the app without providing any personal information.
What we collect
We collect very little, and none of it identifies you personally.
- Question serve counts and feedback. When a question is displayed, the app sends a count increment to our server. If you tap a thumbs up or down on a card, that anonymous vote is added to a per-question tally as well. These are simple aggregate counts — no user identifier, device ID, or location is included.
- Standard server logs. Like any web server, gokubi.com automatically records IP addresses and request timestamps in server logs for security and troubleshooting purposes. These logs are not used for analytics and are not shared.
- Operational log of AI chat calls. If you use the AI chat feature ("Sit with this"), each call writes one short line containing a UTC timestamp, a 16-bit hashed bucket derived from your IP, the token counts of the call, and the conversation length. No message content is recorded. See Operational usage log for the full detail.
- Automatic error reports. If the app encounters a software error, it automatically sends a short diagnostic report to our server. This contains only the error message, a code location (stack trace), the app version, and your browser's user agent string — no chat messages, no question text, no personal information. See Automatic error reports for full detail.
We do not collect: names, email addresses, phone numbers, device identifiers, precise location data, browsing history, behavioral analytics, advertising profiles, or the content of anything you type in "Sit with this".
What's stored on your device
Get Real stores a small amount of data locally in your browser using localStorage (a standard web storage mechanism). This data never leaves your device and is not accessible to us.
- Question cache (
gr-questions-cache-v3) — a copy of the question deck, so the app loads quickly on repeat visits. - Question cache version (
gr-questions-version) — an integer that tells the app whether the cached deck is current; used to refresh the deck when we publish an updated question bank. - UI preferences — whether you've dismissed the swipe hint banner, whether you've granted consent to "Sit with this", per-question timer settings, your expansion-pack selection, your appearance (light/dark) choice, and similar one-time interface states.
- Your thumbs votes (
gr-votes) — which questions you've rated up or down, so the buttons reflect your choice when a card reappears. Only the aggregate tally is sent to our server; this local copy is never tied to your identity.
No cookies are set by Get Real. No session data is stored server-side.
You can clear all locally stored data at any time through the Privacy section of the app menu, or through your browser's settings.
AI chat feature ("Sit with this")
The "Sit with this" feature (formerly called "Play Solo") lets you explore a question one-on-one with an AI companion. This feature transmits data to a third-party AI service. Here is exactly what is sent, by whom, and why.
What data leaves your device when you use "Sit with this":
- The current question text — so the AI knows what you're reflecting on.
- Every message you type — the full text of each message you send during the conversation.
- The AI's previous responses — sent back each turn so the AI can maintain context.
- Your IP address — included automatically by your device in any network request.
Where it goes: Your messages are sent to a proxy server hosted on gokubi.com, which forwards them to Anthropic, PBC (the maker of Claude AI), located in San Francisco, CA. The proxy does not log, store, or inspect the content of your messages — it exists to keep our API credentials private. Anthropic generates a response and sends it back through the same path.
What Anthropic does with your data: Anthropic receives and processes your messages to generate AI responses. Under Anthropic's standard API terms, your messages are not used to train AI models. Anthropic may retain API inputs and outputs for up to 30 days for trust and safety monitoring before deletion. Their full data practices are described in the Anthropic Privacy Policy.
What we store about your chat: We do not log, store, or retain the content of your chat messages on our servers. Once a response is returned, the message text is gone from our infrastructure. The conversation lives only in your device's memory for the duration of your session and is erased when you close the chat or leave the page. We do keep a short, content-free operational log of each call — see Operational usage log below for exactly what it contains.
Sensitive topics. Get Real questions can touch on personal, intimate, or emotionally sensitive subjects. We encourage you to share only what you are comfortable transmitting to a third-party AI service. The AI is a conversational tool, not a licensed therapist, counselor, or medical professional.
The feature requires your explicit consent before first use. You can revoke that consent at any time in the Privacy section of the app menu. The core card game works fully without this feature.
Operational usage log
Every "Sit with this" message costs us a small amount at Anthropic. To keep that cost under control and catch abuse before it becomes a bill, our proxy server writes one short line to a server-side log each time the feature calls the AI. This log is content-free — it records how much the call cost, not what was said.
Exactly what each line contains:
- A timestamp in UTC (e.g.
2026-04-11T21:08:42Z). - A 4-character bucket id — the first four hex characters of a one-way SHA-256 hash of your IP address. This is a 16-bit bucket (roughly 65,000 possible values) that collides heavily across unrelated users, so it does not uniquely identify you or your IP. We use it only to notice if a single bucket is responsible for a disproportionate share of traffic (an abuse signal).
- The number of input and output tokens the call used — small integers that measure how much text was processed. Tokens do not reveal what the text said.
- The number of messages in the conversation at the time of the call.
An additional engagement log for "Sit with this" records, in a separate file, an ephemeral session correlation ID alongside the same timestamp, IP-bucket, and token counts above. The ID is a random 128-bit UUID generated by your browser when you open a chat; it is held only in memory for the duration of the chat overlay and discarded when you close it. It is not stored in cookies, local storage, or anywhere else on your device, and we do not link it to any identity. We use it only to group the turns of one chat into one row of usage when reviewing whether the feature is earning its keep.
What it does not contain: message text, the question, your name, your email, device identifiers, cookies, precise location, or any profile data. We have no way to reconstruct a conversation from this log.
Retention: log files rotate monthly and are deleted automatically about two months after rotation. They are stored on our webserver only, are never shared with third parties, and are not used for analytics, advertising, or profiling.
Why we keep it at all: to monitor the daily cost of the AI chat feature and to detect patterns that look like abuse (e.g. one bucket generating many times the normal call volume). Without a cost signal we would have no way to keep the feature free to use.
Daily-cap counter. To enforce the 30-calls-per-day limit, our server keeps one small database row per calendar day containing a one-way SHA-256 hash of your IP address and a call count. The hash cannot be reversed into your IP address, is never combined with any other data, and rows are deleted automatically after 7 days. The daily spend total — a single number per day with no user data attached — is kept for 90 days.
Guardrail denials log. When the proxy refuses a "Sit with this" request — because the daily budget is exhausted, because your IP has already made 30 calls today (tracked via the one-way hash described above), or because requests are arriving faster than 20 per minute — it writes one additional content-free line to a separate log file. Each line contains a UTC timestamp, which limit was hit (budget, ip_cap, or rate_limit), and the same 4-character IP bucket. This log lets us verify the caps are working and that the client is correctly disabling "Sit with this" when it should. Same retention and sharing rules as above.
Automatic error reports
If the app encounters a JavaScript error during normal use, it automatically sends a short diagnostic report to our server. This helps us find and fix bugs — especially on devices and browsers we cannot test ourselves.
Exactly what each report contains:
- A timestamp in UTC.
- A 4-character bucket id — the same 16-bit IP-derived hash used in the operational usage log. It does not uniquely identify you.
- The app version (e.g.
2026-04-12a). - The page URL path (e.g.
/getreal/index.html) — the page where the error occurred. - The error message and stack trace — technical details about what went wrong in the code. These are generated by the browser, not by you.
- Your browser's user agent string — a standard identifier that describes your browser and operating system (e.g. "Safari 18.2, iOS 18.5").
What it does not contain: the text of any question, any chat message you typed, your name, email, device ID, or any other personal information. Error reports describe code failures, not user activity.
Rate limiting: the app sends at most 10 error reports per page load and the server accepts at most 5 per IP per minute, to prevent any single device from flooding the log.
Retention: error reports are stored in monthly log files on our web server, rotated and deleted on the same schedule as the operational usage log. They are never shared with third parties.
Third-party services
Get Real uses the following third-party service. It has its own privacy policy.
- Anthropic — powers the AI chat feature when you use "Sit with this". Anthropic Privacy Policy.
We do not use advertising networks, marketing trackers, social media pixels, web fonts, or analytics platforms of any kind.
Children
Get Real is intended for users aged 17 and older. The app includes content on topics such as sexuality, relationships, and emotional intimacy that is not appropriate for children. We do not knowingly collect information from anyone under 17.
Your rights
Because we do not collect personal information linked to you, there is generally no personal data for us to provide, correct, or delete. However:
- Device data is fully under your control and can be cleared at any time through the app's Privacy section or your browser settings.
- AI chat messages are not retained by us. For any data retained by Anthropic, please refer to Anthropic's Privacy Policy and their data deletion process.
If you are a California resident, you have rights under the CCPA. Given the minimal data we collect, most CCPA rights are satisfied by default — we do not sell personal information, and we do not have personal information about you to disclose in most cases.
If you are in the European Economic Area, the United Kingdom, or Switzerland, your interactions with Get Real are governed by GDPR. Our lawful basis for any minimal data processing (server logs, count increments, the AI chat operational log) is legitimate interest in operating, securing, and controlling the cost of the service.
Changes to this policy
If we make material changes to this privacy policy, we will update the "Last updated" date at the top of this page. Because Get Real does not collect email addresses, we cannot notify you directly — we encourage you to review this page periodically.
Contact
Questions or concerns about this privacy policy? Reach us at:
Get Real
play.getrealgame.com
info@getrealgame.com
This policy does not constitute legal advice. If you have significant privacy concerns, please consult a qualified attorney familiar with the laws in your jurisdiction.