1. Current beta data

Web search terms, personalization answers, and copied guides are not transmitted to the server. At sign-in, Firebase Authentication authenticates the Google account and supplies a signed Firebase ID token containing a verified email and account-specific ID. The server verifies its signature, audience, issuer, expiry, and Google sign-in provider. AI Konas does not request or store Google access or refresh tokens. The application database stores the Firebase account ID, a non-reversible email hash, account timestamps, usage counts, subscription state, and connection-key hashes; it does not store the email address itself or receive a password. A signed, expiring login Cookie is stored in the browser with security attributes. The hosting provider may process IP address, time, and technical request logs for security and reliability. The browser may store the display language, a recent method ID and date, up to 30 IDs of methods you choose to save, method IDs used to avoid duplicate view-count submissions on the same day, business-pack ID, current step, completion times and next-run date, plus a contribution draft only when you choose to save it. Except for the anonymous aggregate view count described below or when you optionally join improvement measurement, these usage values are not sent to the server. Device values remain until you delete them or clear browser storage. A draft older than 30 days is deleted the next time AI Konas is opened. Do not save on a shared device.

When you submit, the service checks that you are signed in, stores the title, body, original language, authorship type, optional source URL, pseudonymous user ID, publication and certification states, and timestamps, and publishes the post immediately. Only the signed-in author may edit or delete it. AI Konas certification is a separate periodic operator review; only certified posts receive the mark. Deletion removes the public post and related ratings and reviews, although limited records may remain when required for law, safety handling, or backups.

Access counts without personal identification

The library and Loop Studio show an approximate count of browser sessions active in the last few minutes. A random browser ID is irreversibly transformed on the server; the raw ID, name, email, and entered content are not stored. Records expire no more than four minutes after the last heartbeat. This is an approximate session count, not verified people.

Google Analytics loads only when a measurement ID is configured and you explicitly consent. URL queries and fragments, AI Konas user IDs, email, contributions, searches, and entered content are not sent. Ad personalization and Google signals are disabled. Declining does not limit features.

To understand use of public pages, the server aggregates the viewing date, broad page category, broad referrer category, and page-view count by day. Analytics does not issue a Cookie or browser-storage identifier and does not store search terms, URL queries or fragments, contribution or input text, raw IP addresses, user agents, email addresses, personal or device identifiers, or raw referrer URLs in the analytics database. We do not track a person across days, classify new or returning visitors, or use these counts for advertising. Aggregates are retained for no more than 400 days. Because bots, shared networks, and delivery caches may affect the count, it is described as page views rather than people.

Each method card may display how many times its detail view has been opened. The server stores only the day, method ID, method version, and aggregate count; it does not link this count to a browser ID, account, IP address, search term, or entered content. The browser keeps method IDs viewed that day to reduce repeated submissions from the same device. The displayed count is not a count of unique people.

Optional business-pack improvement measurement

Only after you actively select the checkbox while signed in, the server derives a pseudonymous pilot ID from the authenticated account using a secret salt. Pilot events store that ID—not the raw email—plus the business-pack ID, method ID where required, start or completion type, duration and result. After a full run, we also store an optional answer only when you choose “saved at least 15 minutes.” Improvement events do not include input text, search terms, source documents, outputs, names, contact details, customer information, or payment details. Declining, a network failure, or skipping the question never limits features.

You can leave from the screen at any time. Leaving stops new events and, while signed in, sends a request to delete events for your pseudonymous ID. Retention is normally 30 days and cannot be configured beyond 90 days. Hosting-provider technical logs may be processed separately for the period needed for security and reliability.

Post-use star ratings

Ratings work without sign-in. We store a pseudonymous ID derived from a random browser ID with a secret salt, method ID and version, a one-to-five star value, a fixed reason for lower ratings, and timestamps. The raw browser ID, email address, input text, search terms, and AI output are not stored. Written reviews require sign-in and are published only after moderation. One current rating is kept per browser, method, and version and can be changed later. The star average appears after at least five ratings.

Contact requests

The contact form stores the topic, subject, message, reply email, handling status, and timestamps. The reply email is encrypted by the application before storage and is visible only to the authorized operator in the private inbox. This data is used to handle the request, verify identity where needed, address safety issues, and keep legally required records. Do not submit passwords, payment-card numbers, secret keys, or another person's personal data.

2. Data after live launch

Live features may process account details, authentication data, usage and device logs, submitted content, support communications, and subscription status. Passwords will not be stored in plain text. Payment card data will normally be handled by a payment provider.

3. Purposes

• Authentication and account administration
• Search, saving, contribution, update, and AI connection features
• Support, security, fraud prevention, and important notices
• Quality, safety, and discovery improvements
• Billing, contracts, and legally required records

Marketing consent will be optional and separate.

4. Security, processors, and AI connections

We apply appropriate access controls, encryption, permission management, and monitoring. Providers may support authentication, hosting, payments, and analytics under oversight. Personal data will not be disclosed without consent unless permitted by law. Required information or consent will be provided before international transfers. MCP is designed to use only the short information needed to select a method. Because the connected AI initiates the request, review the data and permissions shown by that provider before enabling it. MCP goals are not used for advertising or model training by AI Konas.

5. Rights, retention, and contact

Subject to applicable law, you may request access, correction, suspension, or deletion through the contact form. Identity verification may be required. Data will be kept only as long as needed for stated purposes or legal obligations. The account page supports sign-out.

6. Minors and changes

Minors should obtain guardian consent where required. Material changes will be announced before they apply, and additional consent will be obtained when required.