Last updated: November 11, 2026 (aligned to shipping features; planned integrations clearly marked)
rmmbrly is a private search engine for everything you save online, provided by rmmbrly, Inc. ("rmmbrly," "we," "us"), which is responsible for the data described here. This policy describes what data we collect, how we use it, and your rights as a user.
If you have questions about this policy, contact us at privacy@rmmbrly.io.
When you use rmmbrly, we collect the following:
Photos and screenshots you upload are not listed or published by rmmbrly. The original file is kept in private storage that only your account can read. To show the photo in the app, we also keep a display copy, and we keep a copy of the preview image of links you save, at a web address made from a fingerprint of the image. That address cannot be guessed and is not listed anywhere, but anyone who has that exact link could open it. We are moving these copies to private, short-lived links. Before an uploaded image appears in your library it is screened for objectionable content using Google Cloud Vision SafeSearch. PDF, video, and audio file uploads are not available in the current release. Data-export files used for bulk import are parsed in your browser/device; we store the resulting saved items rather than the original export file. When you delete the item or your account, the original upload is deleted from storage and the display copy is removed by a background clean-up. Each distinct image is stored only once, so a copy stays while another saved item (yours or another user's) still uses exactly the same image.
If you explicitly enable capture-location context in rmmbrly's privacy settings, the app may request your approximate location to help remember where you found something and reconnect saves associated with the same place. Location access is optional and is not required to use rmmbrly.
The device may temporarily observe latitude and longitude in order to create the location cue, but rmmbrly does not persist precise GPS coordinates. Before storage, the server reduces the coordinates to coarse location information and may store an approximate latitude/longitude together with locality, administrative area, country code, and observation time. This coarse location context is associated with your account and may be attached to saves for app functionality and personalization. We do not use location data for advertising or tracking.
You can turn capture-location context off at any time in privacy settings. Turning it off clears the cached location context and stops new saves from receiving location cues.
Not yet available. rmmbrly does not currently sell subscriptions or process payments, and the iOS app ships without in-app purchases. If we introduce paid plans in the future, we will collect your subscription tier and status, and purchase history would be managed by our payment processor (we would not receive or store your card details). We will update this policy and our App Store privacy disclosures before any billing goes live.
The website and, from version 1.1, the iOS app send diagnostic data to Sentry. Error reports are always sent, so we can find and fix what breaks. They contain technical details such as the error message, the screen and feature involved, the app version and build, and the device and operating system, plus a short trail of the screens visited just before the error. Before a report leaves your device, cookies, request bodies, and sign-in headers are stripped, and query strings are removed from the address of the current page and of network requests. The trail of earlier screens removes query strings and fragments and redacts names, handles, and share or invite tokens from app routes. Item identifiers can remain in route paths. Performance timings, and a short replay of the screen around an error (with all text and typed input masked and images blocked), are sent only if you have turned on analytics in Profile settings, and only for a sample of sessions. Diagnostic data is not linked to your account: we do not attach your email, name, or account ID to it. It is listed as Diagnostics, not linked to you, on our App Store privacy label.
We use your data to:
Planned — not yet active. Paid subscriptions and payment processing are not part of the current release. We do not process payments today; if that changes we will update this section first.
We do not use your data to:
We use the following services to operate rmmbrly. They handle data on our behalf, each under its own terms and privacy policy. The Internet Archive and Google Drive rows work differently, as each row says:
| Service | Purpose |
|---|---|
| Supabase | Database, authentication, file storage, edge functions |
| OpenRouter | LLM model routing (for AI enrichment); passes some requests on to other model providers, including Google's Gemini models |
| OpenAI | Text embeddings (semantic search) |
| Anthropic | LLM (some enrichment flows) |
| Microsoft Azure | Backup provider for text embeddings and AI enrichment, used only when our main providers are unavailable. Requests are processed in the United States. |
| Google Cloud Vision | Content moderation (SafeSearch) of uploaded images |
| ElevenLabs | Audio processing |
| Firecrawl | Web page fetching for saved URLs |
| Jina AI (Reader) | Reading the text of a saved web page when our main page fetcher cannot |
| Bright Data | Fetching saved links from sites that block our servers, such as Instagram and Facebook posts |
| Apify | Retrieving the public video, audio, and caption of Instagram, TikTok, and Facebook posts you save |
| Internet Archive (Wayback Machine) | Only if you turn on "Keep public copies on the Internet Archive" in Profile, which is off by default. While it is on, when you save a public web link we may ask the Wayback Machine to archive a snapshot of that page. We send only the link, not your account details, but archived snapshots are public on archive.org. We never send private saves, or links you saved before you turned the setting on. Turning it off stops new requests; snapshots the archive already made stay on archive.org. |
| Google Drive | Reading files from your Drive, only if you connect it (see "Google Drive" below) |
| Sentry | Error and performance monitoring (website and iOS app; see "Diagnostic data" above) |
| PostHog | Product analytics (only after you opt in) |
| Vercel | Web hosting, and server functions that extract the audio and still frames of videos you save |
| Stripe (planned) | Payment processing — not active in the current release |
| RevenueCat (planned) | Subscription management — not active in the current release |
Each of these services has its own privacy policy. Your saved content is sent to AI providers only to generate enrichments (tags, summaries) and to screen content. We do not use it to train models. We have not turned on any provider-specific "zero data retention" setting, so each AI provider keeps and uses the requests we send under its own API terms; requests that go through OpenRouter are handled by whichever model provider OpenRouter sends them to, under that provider's terms.
To read a link you save, our servers request it from the site that hosts it, and for social and video links (such as YouTube, TikTok, X, Reddit, and Instagram) also from that platform's public embed service. Those sites see a request from our servers for that link, not your account details.
Some images in the app load straight from other services, so those services see your device's IP address and the website domain or image involved:
Not yet available — planned. The Instagram account connection described below is not live in the current release. The app does not connect to your Instagram account, request OAuth access, or read anything from your account today. When you save a link to a public Instagram post, our servers read that public post the same way they read any other saved link (see "Third-party processors" above). You can already bring in your Instagram saves by uploading your own Instagram data export — see How you can save below. The section that follows describes how the planned account connection would work if and when it ships; we will update this policy before enabling it.
If you choose to connect your Instagram account, rmmbrly uses the Instagram Login API (operated by Meta Platforms, Inc.) to read posts and reels you have authored. We request only the instagram_business_basic scope. We do not request, read, or store posts authored by other Instagram users on your behalf.
What we receive from Meta when you connect:
How we store and use it:
Disconnecting:
You can choose to connect Google Drive from the Integrations page (in Profile, under Integrations). rmmbrly then asks Google for read-only access to your Drive (the drive.readonly permission; it cannot change or delete your files) and imports new files, or the files in one folder you pick, as saves in your library. The connection token is stored encrypted on our servers and is never sent back to the app. You can disconnect at any time from the same screen, which removes the stored token, or revoke access from your Google account settings. Nothing is read from Drive unless you connect it.
Not yet available — planned. Direct account sync for the services below is not live in the current release. The app does not connect to X, Reddit, or YouTube, hold OAuth tokens for them, or sync from their servers today. As with Instagram, you can already import your saves from these services by uploading their data exports — see How you can save. The section below describes how the planned sync would work; we will update this policy before enabling it.
Beyond Instagram (above), you would be able to optionally connect other accounts so rmmbrly syncs items you have saved on those platforms into your library. Each integration is opt-in — nothing is synced unless you connect it.
| Account | What we read (your own activity only) |
|---|---|
| X / Twitter | Your bookmarked posts |
| Your saved posts | |
| YouTube | Your liked videos (the YouTube Data API does not expose the Watch Later playlist, so we sync liked videos) |
How we store and use it:
You can add items to rmmbrly in several ways. Whichever you use, the resulting item is stored in your library and enriched as described above.
rmmbrly is a logged-in product, not an ad-supported one. We use cookies and browser storage only for the minimum needed to keep you signed in, remember your in-app preferences, and understand which features work. We do not set advertising cookies, third-party tracking cookies, or social-media pixels.
| Cookie / storage | Set by | Purpose | Retention |
|---|---|---|---|
sb-*-auth-token (and related) |
Supabase | Strictly necessary — keeps you signed in across page loads. Without it you would be logged out on every refresh. | Session + refresh window (rotated by Supabase Auth) |
ph_* cookies + localStorage entries |
PostHog | Product analytics — feature-funnel and retention metrics, in our own PostHog instance. First-party; no cross-site tracking. | Up to 1 year (PostHog default); reset when you clear browser data |
theme, library-view, and similar UI preferences |
rmmbrly (localStorage) | Strictly necessary — remembers your dark/light mode and last-used filters so the UI doesn't reset between visits. | Until you clear browser data |
We do not set or allow:
You can clear all rmmbrly cookies and localStorage at any time from your browser's site settings (signing out via Profile → Sign out revokes the auth token server-side as well). On iOS, the app uses native secure storage instead of cookies for the auth session.
Analytics preferences in the public website footer let you choose whether to share optional visit and product-interest actions with rmmbrly. This is off until you explicitly confirm this purpose; an older Profile analytics choice alone does not enable it. We use the same analytics preference as Profile, plus a browser-stored purpose confirmation. Turning analytics off in either place stops new optional measurement.
After confirmation, our first-party telemetry records a random, in-memory session identifier, a versioned interaction rule and allowlisted campaign tags. Product-interest clicks alone do not qualify a need. We qualify a stated saving need only when you affirm the existing saved-backlog checkbox and a new beta request is accepted, with consent already enabled before submission. Duplicate or failed requests do not qualify; raw form answers and identity are not added to visit events. Separately, we count behavioral product interest only when you explicitly explore Product, How Recall Works or Use Cases and then choose a signup or beta-access action within 30 minutes. Hover, elapsed time and page views alone do not qualify. This does not establish your occupation, audience fit or successful use of the product. Reloads, sign-out and withdrawal of analytics consent end session continuity; we do not set a persistent visitor identifier for this measure.
If you sign up in the same consented session, its identifier can associate the visit with your account-creation observation. Missing tags, other devices, App Store journeys and delayed consent can remain unassociated. Once signed in, the existing PostHog analytics choice allows account-linked usage analytics. We do not load PostHog for this anonymous visit measurement. Existing cookieless website analytics remains separate. Visit event payloads contain no email, name, form text or full URL; normal network infrastructure can process connection information such as IP addresses. These raw first-party analytics records follow the existing 90-day cleanup policy; de-identified aggregates may remain longer. Opting out stops future optional events and does not erase previously consented records.
You can:
rmmbrly is available only to adults aged 18 or older. People under 18 are not eligible to create or use an account. If you believe a person under 18 has provided us with data, contact us at privacy@rmmbrly.io.
Our launch policy is 18+ only. Version 1.1 requires an 18+ App Store age rating to match that policy.
We will update this policy when our data practices change. The "Last updated" date at the top reflects the most recent change. We will notify users of material changes via in-app notification or email.
For privacy questions, requests, or complaints: privacy@rmmbrly.io
© 2026 rmmbrly, Inc. All rights reserved.