Privacy Policy - Ownix

Ownix is operated by Leon Eidelman, an individual based in Israel. Leon Eidelman is the controller of the personal information described in this policy. This policy is governed by the laws of the State of Israel. Contact for anything below: leoneidelman09@gmail.com.

• Your Telegram identity (chat ID, name, username, and profile photo URL) - used to identify your account and route your data. Required: Ownix is a Telegram bot, so there is no way to use it without this.
• The email address you provide for invite approval, plus your access status (pending, approved, or blocked). Required to be considered for access while Ownix is invite-only; if you don't provide one your request stays pending and unreviewed.
• Links and documents you save, the analysis Ownix generates from them (transcripts, descriptions, tags, summaries), and your job history - what you submitted and when. Kept until you delete the job or link yourself, or delete your account; see "Deleting your data" and "Deleting your account" below.
• A numeric embedding (a similarity fingerprint, not human-readable text) generated from each saved item's content, to power your own search and link graph. See "Your Brain is private" below for who can see this.
• If you connect your Google account: an OAuth token scoped to Google Drive (drive.file) and Google Sheets (spreadsheets). This lets Ownix create a folder named Ownix and a spreadsheet in your Drive and write your results there. drive.fileonly lets Ownix see or edit files it creates itself, or files you explicitly open or share with Ownix - it cannot access any other file already in your Drive. Ownix currently only creates files; it doesn't use file-picker or "open with" flows to access files you didn't create with it. Optional - Ownix works without it.

Content you submit (a video, article, repo, or document) can itself contain personal information about other people, not just you - a name mentioned in a transcript, a face in a photo. Ownix processes that content the same way it processes anything else you submit; it does not try to detect or specially handle third-party personal data inside it.

Your Index and Second Brain (the link graph on the Brain page) are private to your Telegram account. Nobody else using Ownix can see your saved links, titles, descriptions, tags, or embeddings - every read and write is scoped to your own account. There is currently no feature that shares your saved items with other users.

Google's Gemini API does the actual work: transcribing video, describing images, pulling text off photos (OCR), and writing the summaries, tags, and embeddings you see in your Feed and Brain. Whatever you submit - a URL, a file, a transcript - gets sent to Gemini to make that happen. We use the standard Gemini API (not Google Cloud's Vertex AI); Google's current Gemini API Additional Terms of Service govern how Google may use that data on their end, and we'd rather point you at those live terms than promise something we can't enforce on Google's infrastructure.

Ownix is invite-only while it is young. We use the approval email to review access requests, associate the request with your Telegram account, and contact you about access if needed - this part is not optional while your account is pending.

Ownix is shaped with the people using it. Separately, and only with approved accounts, we may use this email to ask for feedback, discuss feature improvements, and share Ownix product updates. These are manual, one-off emails from a real person - not an automated mailing list, and not yet backed by a suppression list. Tell us at the contact above if you'd rather not get these and we'll do our best to honor it going forward; it never affects your access.

We do not sell your email or share it with third parties.

• We never request access to your Gmail, your existing Drive files, or any Google data beyond the scopes listed above.
• We never sell your data. We send it to the providers below only to get your job done, each scoped to what that step of the job needs.

Turning a saved link into a summary means sending your URL, file, or transcript to outside services, each scoped to what its part of the job needs:

• Google (Gemini API)- transcription, summarization, tagging, OCR, and embeddings, the same model already doing this work everywhere else in Ownix, so your content isn't routed through a second AI vendor just for this.
• Google Drive and Sheets- if you've connected your Google account (see "What we collect" above), Ownix writes results into your own Drive and Sheets - your account, not a shared Ownix datastore, so you keep the copy and the access.
• Google Cloud Storage- separate from the above: Ownix-operated infrastructure, not part of your Google account, used to hold document pipeline artifacts (uploaded PDFs and their extracted text) while a job runs. Removed when you delete the job; see "Deleting your data" below.
• Jina AI - fetching and parsing article pages you submit, rendering JS-heavy pages and stripping out ads and navigation chrome so you get the article, not the page around it.
• Brave Search- extra web search we use while enriching some jobs, run through an independent index rather than Google's, so enrichment isn't tied to the same company already handling your AI processing and storage.
• GitHub- reading repository content for repos you submit, since that's where the repo you linked to actually lives.
• The source platform itself (YouTube, TikTok, Instagram, etc.) - fetching the video/transcript you linked to, the origin of what you saved.

Our use of information from Google Workspace APIs (Drive, Sheets) adheres to the Google API Services User Data Policy, including its Limited Use requirements.

Ownix is operated from Israel, but the providers above (Google, Jina AI, Brave, GitHub, and the platform a link came from) process data on their own infrastructure, which may be outside Israel. Sending your content to them for processing is how Ownix does its job; we don't control where each provider locates its servers.

Video analysis results, job history, links, and embeddings are stored in a private SQLite database scoped to your Telegram account. Document pipeline artifacts live in the Cloud Storage bucket described above. Your Google OAuth token, if you connect one, is stored encrypted and used only to write your own results to your own Drive/Sheets.

Jobs, links, and embeddings: kept until you delete them individually, or delete your account (see below). Your Google OAuth token: kept until you disconnect or delete your account. Your invite-approval email and status: kept while your account exists. Operational logs (request and error logs used for debugging and abuse prevention): kept only as long as needed for that purpose, then rotated out.

Deleting a job or a Brain link pulls it from the dashboard right away - the database record, embedding, transcript, and tags are gone immediately. That same action queues a background cleanup of the underlying cloud files: a job's cleanup covers Drive, Cloud Storage, and Sheets, while a Brain link's cleanup covers its Drive file. That cloud cleanup step is best-effort: on a rare provider-side failure, a cloud artifact can outlive the record you deleted, though the database record itself is already gone. Delete jobs and links one at a time from the dashboard, or delete your whole account at once - see below.

Settings → Danger zone → Delete my account permanently removes, immediately: every job, every Brain link and its embedding, your tags and domain rules, your Chrome extension tokens, your Google connection (the stored token is deleted and revoked with Google), and your account record itself. Your session ends at the same time. Cloud artifacts behind those jobs and links (Drive files, Cloud Storage objects, Sheets rows) are queued for best-effort cleanup the same way individual deletes work, described above. This cannot be undone.

Whether or not you're in Israel, you can ask us (contact above) to:

• Access a copy of the personal information we hold about you - most of it is already visible in your own dashboard (Feed, Brain, Controls).
• Correct information that's wrong or outdated.
• Delete your data - self-serve via Settings → Danger zone, or ask us to do it for you.
• Object to, or ask us to stop, a specific use of your data (e.g. product-update emails - see "How we use approval email" above).
• Withdraw consent for anything we process on that basis, and revoke your Google authorization at any time (see "Revoking access" below).

We'll respond as quickly as we reasonably can - this is currently a one-person operation, so treat any timeline as best-effort rather than a guaranteed SLA.

We use reasonable technical safeguards for the data above: encryption in transit (HTTPS) to every provider listed, your Google OAuth token encrypted at rest, access scoped per Telegram account in the database and API, and Google API credentials restricted to the specific scopes listed in "What we collect."

Ownix is not intended for, and we do not knowingly collect information from, anyone under 18.

Ownix may set a session cookie named ownix_preview to keep the read-only preview active while you browse. It does not identify you or track you across sites.

Send /disconnect to the bot at any time to revoke your Google connection - this deletes your stored token and revokes it with Google. You can also revoke access directly from your Google Account's third-party access settings.

If how Ownix handles your data changes materially - including a new feature that shares data across users - we'll update this page and change the date at the top. For a change to how we use data from Google APIs specifically, we'll notify you and, where required, ask again before the new use begins.

Any questions? Email me here, leon