Privacy Policy
This Privacy Policy explains how G-SYSTEMS GLOBAL INC., doing business as G-SYSTEMS GLOBAL INC. ("Company", "we", "us", or "our"), collects, uses, and protects your information when you use the gFit ChatGPT Plugin (the "App").
It supplements, and does not replace, our general Privacy Policy, Consumer Health Data Privacy Policy and Terms of Use, which govern the gFit website and mobile application. Where this Policy is silent, those documents apply.
1. Information We Collect
The App operates as an MCP (Model Context Protocol) server inside ChatGPT. We collect minimal information required to generate your workout plan:
- Tool inputs: body zones, difficulty, theme (tool get_exercise_catalog) exercise IDs selected by ChatGPT chatSessionId, exerciseIds, exerciseIndex (tool show_exercise). The tools do not request your full conversation, medical history or profile.
- Completion state: When you press Mark as Done, the widget separately saves the highest completed exercise index and its expiry time under a hash of a random workout-session handle in Cloudflare Durable Objects. We do not store the generated workout plan in this completion store. Access requires a separate token delivered to the widget, not the model. The session identifier alone does not grant access.
- Technical data: Application logs contain request method, generic route and response status, not authorization headers, cookies or session handles. IP addresses may be processed in memory to protect the private dashboard login. Cloudflare and our video hosting infrastructure also process network requests needed to deliver the service.
- Quiz transitions: Following the quiz button may increment a daily aggregate count by app source and button placement. This counter stores no IP addresses, cookies, account identifiers or workout-session handles and uses no third-party analytics script.
- App usage: We keep minute-level aggregate counts of tool calls, tool errors, response times, new-plan calls, MCP initializations and widget resource reads, plus coarse client categories when supplied. When ChatGPT supplies an anonymized user or conversation ID, we store that ID as received, without additional hashing, to count distinct recognized users, conversations, per-user new-plan calls and first/last observed use. When supplied by ChatGPT, we also store a country code alongside these usage records to show unique users by country. This optional hint may be inaccurate; missing countries remain unknown. We do not collect cities, regions, coordinates or IP-based location for these analytics. These are pseudonymous identifiers, not account-access credentials. We do not store IP addresses, raw user-agent strings, prompts, exercise selections or tool-result contents in these analytics tables. Missing IDs are not replaced with IP-based fingerprints.
We do NOT collect: personal health records, health metrics, identifiable profiles, names, or email addresses through the ChatGPT Plugin. We do not read your chat history, memory, conversation summaries or files, and we do not use your conversations to train models.
The App requires no account and does not connect to Apple HealthKit or Google Fit. If you also use the gFit app or website, the Consumer Health Data we process there is described in our Consumer Health Data Privacy Policy.
Operational diagnostics record predefined workout filters and body-zone categories, public exercise IDs verified against our catalog (in their selected order), exercise counts, navigation indexes, workflow steps, release versions, error categories, status codes, and timings. Validation diagnostics may include short rejected filter tokens (up to 64 characters), numeric values, known invalid session placeholders such as "null", and duplicate array positions. Actual workout-session identifiers remain excluded. Failed exercise lookups may also record the submitted missing exercise IDs, even when the catalog did not confirm them, to reproduce and investigate the failure. These are limited to three ID-shaped values of up to 200 characters and labeled separately from catalog-verified IDs. Unexpected JavaScript failures may include a shortened exception message (up to 500 characters), with recognizable credentials, contact/network addresses, identifiers, paths and quoted values redacted. A random diagnostic ID connects events of one request; it does not identify an account, chat, or workout session. These records may reveal coarse workout preferences and are not treated as fully anonymous. The video widget may submit a limited playback-failure report containing numeric media states and a one-hour server-signed receipt that links the failure to catalog-verified exercise IDs and the original call. The receipt itself is not logged or stored as a separate user record. Browser runtime exceptions remain local. Custom logs exclude prompt text, full raw tool arguments and results, exercise names and descriptions, session handles, passwords, authorization tokens, cookies, and signed video URLs. Server logs follow the hosting provider's configured retention, separately from completion state.
2. How We Use Information
We use the information solely to:
- Return the requested exercise plans and video tutorials.
- Maintain your session state while you navigate through the "next exercise" flow.
- Monitor the performance and security of the App (e.g., rate limiting).
- Understand aggregate use of the quiz button.
3. Data Sharing and Third Parties
We do not sell your data. We do not share your data with third parties for marketing purposes.
Since the App runs within ChatGPT, OpenAI (the provider of ChatGPT) also processes your prompts and interactions according to their own Privacy Policy. We do not control OpenAI's data practices.
4. Data Retention
Newly saved completion state expires 24 hours after the latest successful Mark as Done save. Reads do not extend retention. Expired progress is no longer returned, and a Durable Object alarm schedules its deletion; physical deletion may be delayed if the infrastructure is unavailable. Starting a new workout does not delete another workout's state. This retention mechanism applies to the new completion store introduced on the effective date above; legacy completion records require a one-time operator cleanup. Application logs and infrastructure logs have separate retention controlled by the hosting provider's configured settings.
App-usage totals, host-provided pseudonymous audience IDs and daily quiz-transition totals are retained for historical usage analysis without automatic expiry. Legacy hourly totals contain no user identifiers. Audience IDs are not linked to the generated workout plan or the completion-progress store.
5. Your Rights
The App does not require an account. Audience analytics uses host-provided pseudonymous IDs rather than names or contact details. If you have questions about your privacy or wish to inquire about data practices, please contact us.
European Union (GDPR). You have the right to access, correct, update, or request deletion of your personal information, to object to processing, and to data portability. Where we transfer personal data outside the European Economic Area, including to the United States, we rely on Standard Contractual Clauses approved by the European Commission or another valid transfer mechanism.
California (CCPA/CPRA). You have the right to know what personal information we collect, the right to request deletion, the right not to be discriminated against for exercising these rights, and the right to limit our use and disclosure of Sensitive Personal Information to what is reasonably necessary to provide the Service.
To exercise any of these rights, contact us at support@gfit.world.
6. Children’s Privacy
The Service is not intended for individuals under the age of 13 (or under 16 in the EEA). We do not knowingly collect personal information from children. If we become aware that we have collected personal data from a child without parental consent, we will take steps to remove that information.
7. Governing Law
This Privacy Policy and any disputes related to it are governed by and construed in accordance with the laws of the State of Delaware, United States, without regard to its conflict of law provisions.
8. Changes to This Policy
We may update this Privacy Policy from time to time. We will post the revised policy here with an updated effective date.
9. Contact Us
If you have any questions about this Privacy Policy, please contact us at:
G-Systems Global Inc. (d/b/a gFit)
Email: support@gfit.world
Website: https://gfit.world/
Address: 131 Continental Dr, Suite 3052, Newark, DE 19713, United States