Effective date: October 21, 2026
Ginupy S.A.S. ("we", "our", or "us"), based in Cuenca, Azuay, Ecuador, provides the Ginupy website and Ginupy Code. This policy explains the information handled by our website, app features, and purchase verification service. The information involved depends on the features and services you use.
1. Website, contact, and browser storage
Our contact form sends your name, email address, message, and an anti-spam verification response to our contact endpoint at ginupy.com. We use information you submit to respond to your inquiry. Please do not include API keys, passwords, private keys, purchase tokens, or confidential source code in a message or public community post.
Our website stores your selected language in your browser’s local storage. The contact form uses session storage to preserve the language of its result page. These are browser storage features, rather than advertising cookies, and can be cleared through your browser settings.
The website loads fonts from Google Fonts. Contact pages load Google reCAPTCHA to help prevent spam. These requests expose technical information, including your IP address and browser request information, to Google. reCAPTCHA may use cookies and other signals under Google’s policies. You can contact us by email if you prefer not to use the form.
Requests to our website and APIs also make network information, including an IP address, available to the servers handling them. The purchase API uses IP-based rate limits for abuse prevention. Website access-log fields, analytics processing, hosting providers, and their retention periods are not specified here as verified practices; please contact us for current operational details.
2. Local AI and downloaded models
Ginupy Code offers a local-model mode designed to run downloaded language models on your device. Local inference is distinct from sending a request to a cloud AI provider. Model discovery and downloading are network activities: the model browser queries Hugging Face, and downloads request model files from their hosting service.
The app provides chat threads and model storage. Local-model mode should not be understood as a promise that every feature of the app is offline: downloads, purchases, remote development, linked content, and connected AI features can involve external services. Only the selected local inference operation is described as local; this policy does not make a blanket guarantee that all chat history, tool activity, or diagnostics remain on the device.
3. Cloud AI, API keys, and selected providers
Connected AI features let you configure a supported provider or gateway and use your own API key. The provider settings allow keys to be saved for later use and checked against the selected provider. API keys authorize use of your provider account; protect them and revoke them with the provider if they are exposed.
When you use connected AI, the request content must be processed by the selected service. This can include the prompt and any conversation context, code, files, or tool results included in that request. A gateway may forward requests to another model provider. Custom endpoints are controlled by their operators. Review the destination and its terms before sending confidential information.
Provider choices shown in the app include OpenAI, Anthropic, Gemini, Grok, OpenRouter, Vercel AI Gateway, Cloudflare AI Gateway, Kilo Code, LiteLLM, and custom endpoints. Listing a provider does not mean it receives data when you have not selected or used it. Provider retention, training use, security, and processing locations depend on the service, account settings, and applicable agreement. We do not guarantee that every provider deletes requests immediately or excludes them from training.
4. SSH, SFTP, files, and source code
SSH connection profiles contain information such as the server address, port, username, remote path, and host-key settings. The app saves these profiles through its settings system. When you choose to save an SSH password or private key, the credential service uses the device secure-storage integration. Unsaved credentials are held in the app’s session memory. The credential service keeps private-key passphrases in session memory rather than writing them to secure storage or app settings.
The app passes the selected connection settings and credentials to its SSH plugin to connect to your chosen server. SSH terminal activity and SFTP file transfers involve that server. The reviewed connection service contains no upload of these credentials to the Ginupy purchase API. This does not describe the logging or storage practices of a remote server you choose.
Ginupy Code supports local workspaces, downloaded models, and chat-thread storage. Files can also be stored on remote servers or repositories you use. Sending content to a remote server through SFTP, pushing to a Git remote, including content in a connected AI request, sharing or exporting it, or running code that makes network requests can transfer information outside the device. The recipient’s practices apply to those transfers.
GitHub sign-in is configured to use a Ginupy-hosted OAuth endpoint. Do not assume that every Git-related authentication step bypasses Ginupy infrastructure. The authorization scopes and handling of credentials depend on that integration; contact us if you need details before connecting an account.
5. Purchases and subscription verification
Apple App Store and Google Play process in-app purchases under their own terms. Our purchase verification interface uses store-issued purchase evidence, not a payment-card entry form. The app sends an Apple signed transaction or a Google Play purchase token, along with the product identifier, app identifier, device identifier, and purchase environment, to the Ginupy purchase service at api.ginupy.com.
The service validates purchase evidence and issues a short-lived authorization token for subscription verification and authenticated error reporting. It also processes store subscription notifications. Server records include customer and subscription identifiers, store transaction identifiers or purchase tokens, product and platform information, subscription status and dates, renewal information, verification results, and relevant store notification data. Device records associate a device identifier with a customer, platform, and last-seen time. A device identifier is not necessarily an advertising identifier.
Store responses can contain additional purchase metadata. Where Google supplies a profile name through its subscription API, the backend can save it as the customer name. Purchase evidence and identifiers are used to provide and restore paid access, reconcile store notifications, investigate verification failures, and prevent misuse. They should not be treated as anonymous merely because they do not include a card number.
6. Usage information, diagnostics, and feedback
The app’s usage screen displays AI usage, token counts, performance measurements, and estimated costs from chat-related data, and offers a sharing action. This feature is distinct from a website tracking or advertising service.
The app includes an error-report setting and an authenticated error-reporting function. A report can contain error text, device type, total and available memory, platform, app identifier, and device identifier. The server associates accepted reports with the authenticated purchase customer. The reporting code removes several common token and receipt formats, but that filtering is not a guarantee that all personal information is removed. Avoid including secrets or sensitive content in diagnostic material you send us.
Beta recruitment, participation guidance, and feedback take place on Discord. Information you post there is processed by Discord and may be visible to other community members. TestFlight and Google Play may also process test feedback or diagnostics under their own settings and policies.
7. External services and disclosure
We do not sell personal information. We disclose information when needed to operate a feature you use, respond to your request, comply with applicable law, or protect our services and users. The selected provider, model host, repository host, remote server, app store, and community service each have their own data practices. Processing may occur outside your country.
Relevant privacy information is available from Apple, Google (including Google Play, reCAPTCHA, and YouTube), Hugging Face, and GitHub. The app also includes YouTube learning-video integration; using it connects to YouTube.
For connected AI, consult your selected provider’s policies, including OpenAI, Anthropic, Google, xAI, OpenRouter, Vercel, and Cloudflare. For other gateways and custom endpoints, check the operator’s policy.
Our links to Discord, X, and Substack open external services. The footer links themselves do not embed those services’ tracking widgets.
8. Retention and security
Different records have different lifetimes. The purchase service is configured to expire access tokens after 60 minutes and schedule cleanup hourly. Error logs and webhook-failure logs become eligible for cleanup after 7 days; device-history records become eligible after one month without activity. Verification history is normally kept for 3 months after the latest attempt, extended to at least 6 months after a failure. Records pending manual review are excluded from that automatic verification-history cleanup.
These rules are not a promise that every purchase record is erased after 7 days. Customer and subscription records, retained store metadata, webhook deduplication records, support correspondence, server logs, and backups have separate purposes. A single fixed retention period for those categories has not been established in this policy. Contact us for the retention applicable to your records. Scheduled cleanup also depends on the service’s operational configuration.
The purchase service stores hashes of issued authorization tokens and encrypts certain verification-history payloads. SSH credentials selected for saving use a secure-storage integration. These protections do not mean every app setting, file, log, or database field is encrypted. No transmission or storage system can guarantee absolute security.
9. Your choices and rights
Depending on the law applicable to you, you may have rights to access, correct, delete, or obtain a copy of your personal information, restrict or object to processing, withdraw consent where processing relies on consent, or raise a concern with a competent data protection authority. Contact us at [email protected] to make a request. We may need to verify your identity and identify the relevant purchase or records. Do not email full purchase tokens or secrets.
You can use the app’s available file, model, chat, and credential controls to manage information on your device, clear website storage through your browser, and revoke API keys with their providers. Removing local app data does not itself delete server-side purchase records, remote repositories, remote files, or third-party records.
The purchase backend provides a verified deletion-request process. An approved request can suspend associated access and schedule permanent deletion after a configurable cancellation period, with notifications sent by email. Some audit status and timestamps remain after the personal request details are erased. Deleting Ginupy records does not cancel an Apple or Google subscription; manage billing through the relevant store. Ask us about your applicable deletion timetable and any records that must remain.
10. Children, updates, and contact
Our services are not directed to children under 13. We do not knowingly collect personal information from children. If you believe a child has provided personal information, please contact us.
We may update this policy as our services change. Updates will appear on this page with a revised effective date. This policy is provided in English, including when the surrounding website navigation is displayed in another language.
- Company: Ginupy S.A.S.
- Location: Cuenca, Azuay, Ecuador
- Email: [email protected]
- Website: https://ginupy.com/