Privacy Policy

RepoFlow is operated by Thomas Deng as an individual. This policy explains data used during anonymous access, optional account sign-in, waitlist enrollment, and feedback.

Data collected

RepoFlow does not require an account email, password, or public profile. It stores a random anonymous authentication identifier, the start and end of the free-access period, selected project-purpose interests, one self-described role, recommendation settings, saved repositories, interface settings, and limited product events such as impressions, qualified dwell, opens, and save changes when personalized recommendations are enabled. If you join the waitlist, RepoFlow stores the email address you enter. If you send feedback, RepoFlow stores the required message, a submission identifier, and a contact email only when you choose to provide one. Waitlist and feedback contact addresses are submitted separately from account sign-in and are not verified by those forms.

Optional account sign-in

Email or Google sign-in creates or restores a RepoFlow account through Supabase Auth. RepoFlow stores the account identity and provider-supplied name, avatar, and confirmed email when available, together with that account’s Saved projects and preferences. Signing in replaces the anonymous browser session; anonymous data is not automatically merged into the account. Account state can be restored by signing in on another device.

How data is used

Data is used to isolate the anonymous browser session, enforce its access period, restore settings and saved items, operate optional personalized recommendations, record waitlist interest, review product feedback, prevent abuse, diagnose failures, and maintain service security. Raw mode disables recommendation-training event collection while keeping Saved available during the access period. RepoFlow does not sell personal information.

Audience analytics

When enabled on the production product, RepoFlow uses Google Analytics 4 to measure aggregate traffic. RepoFlow sends a fixed app-entry event and page-view events carrying a fixed screen label from a closed list (for example /discover/rising, /saved, /repository), the referring website origin, approved public campaign codes, a timestamp, and the ordinary browser and device information a web request carries. The screen label never contains a repository you opened, your search text, or anything describing you. Full incoming URLs and referring paths, arbitrary campaign values, ad click identifiers, query strings, and fragments are excluded. App entry is counted separately from screen views so loading alone does not imply successful product use. RepoFlow does not send its Supabase subject identifier, contact email, waitlist status, feedback text, search text, saved repositories, recommendation events, form contents, or administration-console activity to Google. Google Signals, advertising personalisation, and remarketing are disabled, so this measurement is not used to build advertising audiences. Unlike RepoFlow’s earlier plan for cookieless measurement, Google Analytics sets its own cookies in your browser and sends this data to Google. RepoFlow configures those cookies to expire with the browser session, but they are cookies, and Google processes the data it receives — including your IP address — under Google’s own terms and privacy practices, which RepoFlow does not control.

This audience data is used to estimate visitors, page views, sessions, engagement, referring sources, and browser or device mix. These counts are approximate and may omit visits because of Do Not Track or Global Privacy Control signals, content blockers, network failures, or ordinary measurement limits. RepoFlow disables analytics entirely when your browser sends a Do Not Track or Global Privacy Control signal. Raw recommendation mode disables recommendation-training events; it does not by itself disable this separate aggregate audience measurement. Analytics is disabled on /admin, on static legal pages, in local development, in device preview, and on any production hostname the operator has not explicitly approved.

Audience analytics is active on repoflow.ca, subject to the privacy signals and exclusions described above. Because audience events are not linked to a RepoFlow account or contact identifier, the operator generally cannot locate a particular analytics record from an email address. You can block this measurement with a content blocker or a browser privacy signal without losing any RepoFlow functionality. Visiting /?analytics=off also disables this measurement for this browser using a local opt-out preference; /?analytics=on clears that preference but does not override a browser privacy signal.

Who can access submitted data

The individual RepoFlow operator may view anonymous identifiers, access-period timestamps, waitlist email addresses, feedback messages, and any optional feedback contact email through a restricted administration console in order to operate the service, review requests, respond to privacy requests, and improve RepoFlow. This information is not public.

Service providers and public sources

Supabase provides authentication, database, and server functions and stores waitlist and feedback records. OpenAI Sites provides application hosting. GitHub supplies public repository information. Google provides the audience analytics described above through Google Analytics 4. These providers process data under their own terms and privacy practices.

Retention and control

Anonymous-session, waitlist, and feedback records are retained only as reasonably needed to operate, improve, and secure RepoFlow and to meet legal obligations. Anonymous product state cannot be recovered on another device and may become inaccessible if browser storage is cleared. You may change recommendation settings during active access. Waitlist and feedback contact records may be retained independently after an anonymous subject is removed so the operator can preserve and answer the submitted request. To request access, correction, export, removal from the waitlist, or deletion of feedback or other personal data, contact the address below.

Do not include passwords, credentials, confidential information, or sensitive personal information in feedback.

Security and international processing

RepoFlow uses tenant access controls, managed authentication, encrypted network transport, restricted server credentials, and a separately authenticated, verified, allowlisted operator account. Browser roles have no direct access to administration, waitlist, or feedback tables. No system can guarantee absolute security. Service providers may process data in countries outside your own.

Changes

Material changes will be reflected by updating this page and its effective date.

Contact

Privacy requests, waitlist removal, and feedback deletion: dengmaizi0802@gmail.com.