Privacy Policy

Readster's business model is subscriptions — not ads, not data. This policy is short because the honest answer to most privacy questions about Readster is "we don't do that."

The three commitments

1. We never sell your data, and we run no advertising. There is no ad targeting, no data brokering, no "anonymized insights" product.
2. We never train AI models on your content, and we only work with AI vendors under terms that prohibit training on your data. Your ratings and reading activity are used for exactly one thing: ranking your own brief.
3. Your watchlist, notes, sources, and archive are private to your account. A watchlist with notes can reveal what you're working on — a research direction, an unannounced project, a position. We treat it that way. No one else can see it, and we never share it except as described under "Subprocessors" below.

What we collect

• Account data — email address, name if you provide one, sign-in records.
• The content you connect — feed items, newsletters sent to your private @in.readster.ai address, filings, podcast and video metadata, podcast transcripts and YouTube captions, Bluesky posts. This is the product working, not surveillance: we process it to build your briefs and archive.
• Your configuration — watchlist entries and notes, source list, preferences, item ratings.
• Billing data — handled by Stripe. We never see or store full card numbers.
• Operational data — logs and delivery/bounce records needed to run and secure the service.
• Site analytics — on our public pages only (the marketing site, pricing, blog, legal pages, sign-in and sign-up, and the public filing pages; never the app or your briefs), we count visits with umidata, a cookieless analytics service: the page address without its query string, the referring site, browser, operating system, device type, screen size, language, and a country derived from the IP address. It sets no cookies, builds no profile across sites, and doesn't count your visit when your browser sends Do Not Track.

We do not collect your browsing history, contacts, or anything from devices beyond what the product needs.

Cookies

We use cookies that are required to sign you in and remember your preferences. We do not use advertising or cross-site tracking cookies. Site analytics uses no cookies at all.

How your content is processed

Your content is processed by software — including AI models — to deduplicate, summarize, rank, and cite it in your private brief. Summaries are delivered only to you. We never publish AI-generated summaries of your sources publicly, and human access to your content by our staff is limited to debugging with your consent or as required by law.

Public filing pages. Separately from any account, Readster publishes AI-generated summaries of public-record documents, such as SEC filings, on public pages. These pages contain no reader data: which companies and filings appear is decided by a fixed list, never by what any reader follows, and no page shows a reader's name, email address, watchlist, sources or activity.

On Readster Free

When a trial or a subscription ends, your account moves to Readster Free. Everything in this policy still applies, with these differences in what we process for you:

• We keep your account, configuration, archive and past briefs, and keep sending your weekly brief — and at most one SEC filing alert a day — to your account email address.
• Newsletters that reach your private address are not parsed, summarized or added to your archive. Our email provider (Cloudflare) still receives and stores the raw message, as it does for every account.
• We don't transcribe podcast episodes or read YouTube captions for you. Shows and channels you followed stay on your source list.

Subprocessors

These vendors process data on our behalf. Every vendor that touches your content is bound by terms that prohibit using it to train their models, and every AI request goes through OpenRouter with data collection denied, so it is served only by providers that neither keep your content nor train on it.

• InterServer — hosts the server, in Los Angeles, United States, that runs the Readster application and its PostgreSQL database. Stores your account data and archive.
• Cloudflare — DNS, and the network proxy in front of our website and app: all web traffic passes through it, and it processes requests and pages in transit. Also email sending and routing, including inbound newsletter mail. Sees newsletters sent to your private address (it receives and stores the raw message), outbound briefs and the other emails we send you.
• OpenRouter — routes every AI request below to a provider serving that model, restricted to providers that neither retain prompts nor train on them. Sees everything sent to those models.
• Anthropic (Claude, via OpenRouter — served by Anthropic or by Amazon Web Services) — brief synthesis, and answers to Ask and other on-demand requests about your items. Sees the items being summarized, your watchlist with its notes, your first name and time zone, and your questions.
• DeepSeek (open models, via OpenRouter — served by hosting providers such as Together AI, never DeepSeek's own service) — item summaries and enrichment (the people, companies and topics an item mentions). Sees the content being summarized, including newsletters sent to your private address and podcast transcripts.
• OpenAI (via OpenRouter) — embeddings for ranking and search. Sees content being indexed, your watchlist entries with their notes, and your search queries.
• Google (Gemini, via OpenRouter) — podcast transcription. Sees the audio of episodes you follow.
• Stripe — payments, when you pay for Readster Pro. Sees billing details (never stored by us).
• umidata — cookieless analytics for our public pages. Sees standard request data from visits to those pages; never your account or content.

YouTube videos are read from their captions, which we fetch from YouTube itself; YouTube sees a request for the video, never your account or watchlist.

We'll update this list before adding a vendor that touches your content.

Retention and deletion

• Everything is retained while you have an account — on a trial, on Readster Pro or on Readster Free. The archive is the product.
• Cancel Pro, and nothing is deleted: your account moves to Readster Free and your data stays exportable. If an ended subscription leaves an account on no plan at all (for example, because its email address was never verified), its data stays exportable for 90 days, then we delete its archive. See the Data Export Policy.
• Ask, and we delete. Delete your account in your settings, or email [email protected] from your account address to request deletion or a copy of your data (access, correction, and portability requests work the same way). We honor these regardless of where you live.
• Deleting your account removes your content from our systems, including shared processing caches where you were the only contributor. Backups age out on a fixed schedule after deletion.

Sharing beyond subprocessors

Only if the law requires it (we'll tell you unless legally barred), or to protect the service and its users from abuse. That's the whole list. A change of ownership would bind the new owner to this policy for existing data.

Security

Data is encrypted in transit and at rest. Access is limited and logged. If a breach affects your data, we will notify you promptly and plainly.

Beta status and changes

Readster is in beta; this policy still binds us in full. If we change it materially, we email account holders — on every plan — before the change takes effect and update the date above. We won't weaken the three commitments at the top for data already collected.

Contact

[email protected]