Privacy policy

Last updated: July 23, 2026

The short version: Repubbed stores the content you import so you can search it. Your data lives in your own private space, we don’t run analytics cookies, we never sell anything, and you can export or delete everything yourself at any time.

What we store

When you use Repubbed, we store:

  • Your account basics — the email address you sign in with, an optional display name, and settings like your preferred view.
  • The content you import — titles, bodies, descriptions, tags, dates, source links, and metadata for the blog posts, videos, emails, and social posts you add to your archive.
  • Mirrored images — copies of images referenced by imported content, so your archive keeps working if the original host removes them.
  • Import history — records of your import jobs (what was fetched, what succeeded, what failed) so imports can resume and report honestly.
  • Feedback you send — messages submitted through the in-app feedback form, along with basic context (your account email and the page you sent it from) so we can respond.

Everything above is stored under your own user ID and is readable only by your signed-in account. There are no public profiles and no sharing features.

Where it lives

Repubbed runs on Google Firebase (Firestore for data, Cloud Storage for images, Firebase Authentication for sign-in). Firebase acts as our data sub-processor; Google’s own privacy commitments for Firebase apply to data stored there. The app itself is hosted on Vercel.

Sign-in emails

Repubbed uses passwordless magic-link sign-in. When you enter your email, Firebase Authentication sends the sign-in link — we never see or store a password. Transactional emails from Repubbed itself (for example, a note that a long import finished) are sent through our own email gateway to the address on your account, and only in response to things you did in the app. No marketing lists, no newsletters you didn’t ask for.

Cookies and analytics

Repubbed sets no analytics or advertising cookies. The only browser storage used is what Firebase Authentication needs to keep you signed in, plus local preferences (like your default view). We do not run third-party trackers.

AI features

Repubbed’s AI features are bring-your-own-key: they only work after you connect your own OpenRouter key in Settings, and nothing is ever sent to an AI provider until you use an AI feature yourself.

  • What gets sent — for the content involved in the action you take: the title, excerpt, up to 8,000 characters of body text, and your category and tag names. If you build the AI search index, the text of every item you consent to indexing is sent.
  • To whom — to OpenRouter, and through it to the upstream provider that serves the model you chose (for example OpenAI, Anthropic, or Google).
  • When — only after you connect a key, and only when you trigger an AI action. Nothing runs automatically unless you turn a preference on, and building the search index always asks for explicit confirmation first.
  • Your key — stored encrypted at rest, used only server-side to make the calls you request, and never shown back to anyone (including you) after connecting. You can disconnect it in Settings and revoke it in your OpenRouter dashboard at any time.
  • Storage and retention — OpenRouter states that it does not store your prompts or your content by default; it keeps request metadata (timestamps, model used, token counts) for billing. Do not enable “prompt logging” in your OpenRouter account settings — OpenRouter offers a small discount for it, but its terms grant OpenRouter broad commercial rights over logged prompts. The upstream provider that serves your chosen model has its own retention policy (many retain content briefly — commonly up to 30 days — for abuse monitoring); each model’s provider policy is shown on its openrouter.ai page.
  • Training — whether a provider may train on your content depends on that provider’s policy, not ours. Your OpenRouter account has a privacy setting to refuse routing to providers that train on data — we recommend turning it on. We can’t independently guarantee any upstream provider’s behavior; if that matters to you, choose a model whose provider policy you trust — the model choice is entirely yours.
  • What we store — after indexing, Repubbed stores only the resulting embedding vectors (lists of numbers) in your archive’s database, alongside the content you already keep there. Your content is not stored anywhere new.

Export and deletion

Your archive is yours:

  • Export — Settings → Your data downloads your entire archive (content and taxonomy) as one JSON file, any time.
  • Delete content — you can delete individual items, bulk selections, or your whole archive. Deletes are permanent — there is no trash. Export first if you’re unsure.
  • Delete your account — Settings → Danger zone removes your content, your images, and your sign-in account itself.

Contact

Questions about this policy or your data: use the feedback form in the app, or email pete@petemcpherson.com.