Privacy | Gedighant

Information used for an order

We use a verified email address to create and sign in to your private account, enforce the one-free-image entitlement, and send private result or order notifications. We may also send at most one reminder email about an order you started, such as a video checkout you didn’t finish or a free image you haven’t turned into a video yet; every reminder has an unsubscribe link, and unsubscribing stops all reminders to that address. We use the uploaded source image only for the explicit generation you request, together with the adult, consent, source-rights, and terms attestations you make. We record the applicable attestation versions and acceptance time.

Private storage and access

Uploaded images and generated results are kept in private object storage rather than a public gallery. Order status and result access use an unguessable private order link. Keep that link private.

Payment information

The personalized free image does not require a card, wallet, or blockchain payment. A one-time returning creator image is a paid image offer shown only after the free image entitlement has already been used; its exact price is shown at checkout before payment. If you later choose a paid video, launch payments are settled in USDC on Base. Gedighant records the payment identifiers, status, amount, and public-blockchain evidence needed to reconcile an order and any approved refund. Gedighant does not need your credit-card number for this Base USDC flow.

Embedded provenance

Paid video downloads contain a non-personal embedded provenance marker that Gedighant can use to associate that delivered copy with an internal generation record. The personalized free image is delivered without that video-only marker. The marker does not contain your email, private order token, or another directly identifying customer value. It is an abuse-deterrence and investigation aid, not a promise that the marker will survive deliberate editing or re-encoding.

Retention

Our launch defaults intentionally keep intimate media for short periods. For the account-gated personalized free image, the uploaded free source expires after 12 hours and the free generated result expires 12 hours after completion; automated retention removes those private media objects after that window. For the one-time returning creator image, an abandoned returning-image checkout with no payment attempt or provider generation is released after 12 hours so the offer can be retried, and the returning creator result expires 12 hours after completion. For a paid or abandoned video order, the existing source-image default is 7 days when no live checkout, payment attempt, or generation job still requires it, and paid generated results are deleted 30 days after completion. The email stored on a final order is anonymized after 90 days unless an active removal case still requires it. The verified account email and hashed session records are retained while the private account remains active so Gedighant can authenticate the account and enforce the free entitlement. Contact, signature, and locator information from a closed takedown report is anonymized after 24 months. Minimal authorization, payment, entitlement, and audit records may be kept for up to 7 years for transaction integrity, abuse prevention, dispute handling, exact-once safety, and legal/accounting purposes.

Removal and takedown

A customer can request removal from the private order page. A depicted person or authorized representative can submit a good-faith nonconsent report through the public takedown form without an account. When a report is matched to an order, access can be blocked while deletion is completed.

Service providers and operational records

Gedighant uses infrastructure providers for hosting, database and queue processing, private object storage, Base USDC payment infrastructure, email delivery, personalized-video generation, and automated content-safety screening. Uploaded photos and generated results are sent to our safety-screening provider (Hive) only to check them for prohibited content before anything is created or delivered, and we keep only the pass or block outcome. Hive handles the copies it receives under its own retention terms. We keep the minimal operational and audit records needed to operate the pilot, prevent duplicate paid actions, investigate failures, enforce removal requests, and measure service reliability. To keep the free image to one per person, we store a random device identifier in a cookie and compare it (as a hash), your email address and a fingerprint (hash) of your photo with earlier free-image claims. If your browser hits an error on our site, we record a redacted error report (no email, photo, order link or other identifier) so we can fix it. To understand which public pages bring visitors into the creation flow, we record a few anonymous visit steps, including a separate anonymous page-group view for allowlisted public pages. It contains a random tab identifier and broad first-page and current-page categories, such as home, create, guide or motion; guide and motion slugs, page addresses, referrers, and user-uploaded data are not included. Separate funnel steps record the creation actions and a general photo-fit rating where one applies (never the photo), along with campaign tags (such as utm_source) after checking them against a short, safe format. Private account, payment, order, and takedown pages are not included in page-group views. If you place an order, the broad first-page category and campaign tags are recorded with it too, so we can compare aggregate visits and outcomes by source and page category without connecting anonymous visit identifiers to orders. These first-party steps do not record the page address or the site you came from, and private order pages are never counted this way. They are kept like our other audit records. Separately, Vercel Web Analytics receives anonymous pageview data only for an allowlist of public pages. We normalize individual guide and motion pages to their category, remove query strings, and exclude account, order, payment, takedown and API pages. Vercel may also receive the address of the site that linked you to us, as your browser reports it; when you move between our own pages, your browser shares only gedighant.com, never a page address or link. This lets us see aggregate visitor and page counts without recording private order links or individual guide and motion choices. When our region check turns a request away, we only add one to a daily count of the area it came from (country, or US state), the kind of page, and whether it was a link preview or crawler; nothing else about the request is kept, and the counts are deleted after 35 days. If you arrive through a partner’s referral link, your browser remembers the partner’s code for 30 days and we record it with your account when you first verify your email, so the partner can be paid a share of your purchases. Partners never see your email, photos or creations.

Optional product analytics

If we enable our self-hosted Umami analytics, it receives pageviews only for allowlisted public pages. On the homepage and creation page, Umami receives generic funnel event names, with no event properties or user-provided details. We send a normalized page address without its query or fragment, and do not send page titles, referrers, campaign tags, account or order identifiers, email addresses, wallet or payment information, prompts, filenames, or media. Umami does not use cookies; it uses the request IP address, browser information, and website identifier to generate an anonymous visitor session hash and may derive location metrics from the IP address. Umami states that it does not store the IP address. Our hosting provider may process connection data to deliver the service. Session replay and heatmap collection are disabled. Self-hosted Umami retains analytics data indefinitely by default unless we manually delete it.

Safety and abuse prevention

Requests may be rate-limited and operational events may be recorded to protect the service, enforce capacity controls, and investigate abuse. Media subject to a completed removal request is not restored by an automated retry.