1. Our approach
Your ideas are sensitive. Reports are private by default — nothing is public unless you explicitly share it.
2. Data we collect
- Account data (email) — via our auth provider
- Founder profile (skills, goals, preferences)
- Ideas & analysis inputs (idea text, pasted evidence, optional URLs)
- Reports we generate for you
- Billing metadata — via Polar
- Product analytics: a pseudonymous browser ID (with consent) or account ID, query-free page path, coarse browser/device metadata, event name, and bounded IDs, statuses, counts, booleans, credit amounts and USD values. Never your idea text, pasted evidence, founder profile text or search-query text.
- If you accept analytics cookies: a normalized public landing page, campaign tokens, referrer hostname and the last signup-directed content CTA. We never store the search query, full referrer URL or CTA text in this attribution record.
3. How we use it
To run analyses, deliver reports, operate your account and billing, and improve the product. We do not sell your data.
4. AI processing and search
To generate reports, your idea inputs are sent to our AI providers (Anthropic, and in some configurations OpenAI). They do not train on this data. Your email and billing data are never sent to the AI.
Discovery and First-Customer Leads search the public web on your behalf. Discovery includes the market or niche you type; First-Customer Leads builds queries from the buyer and pain fields of the selected idea. Depending on configuration, those phrases go to one or more of our public-source connectors: Serper or Tavily; Hacker News search via Algolia; Stack Exchange; Remotive; and direct searches on TrustMRR, SaaSHub, or BetaList. Reddit's API is used only after separate commercial approval and configuration. Pages returned by these searches are fetched by us when allowed. Your founder profile, email and billing data are never sent to a public-source connector.
5. Service providers
- Clerk — authentication
- Anthropic — AI analysis
- OpenAI — AI analysis (fallback configurations)
- Public-source services — Serper or Tavily; Algolia's Hacker News search; Stack Exchange; Remotive; and direct listing searches on TrustMRR, SaaSHub, or BetaList. They receive market/niche or buyer-and-pain query phrases when their connector is enabled. Reddit is disabled unless separately approved and configured.
- Polar — payments (Merchant of Record)
- PostHog — product analytics, no idea content. Its browser script loads only after you accept cookies (manage via “Cookie preferences” in the footer). We also record a small number of server-side events — that you signed up, that a run finished, that a purchase completed — keyed to your account and never containing your idea text. Those are not cookie-based and are sent regardless of your cookie choice.
- Sentry — error monitoring
- Neon — database hosting
- Resend — transactional email
6. Sharing
Share links are opt-in, random-token, noindex, and revocable anytime. Recipients see the report only — never your profile, email, or other ideas.
7. Retention
Active data is kept until you delete it. Deleted projects, reports, and accounts are hard-deleted within 30 days.
8. Your rights
You can export all your data (ZIP) and delete your account at any time. For privacy requests: support@whittleos.com.
9. Cookies
Essential cookies keep sign-in and your cookie choice working. If you accept analytics, a first-party attribution cookie is kept for up to 90 days so a later signup can still be connected to the public page and signup CTA that led to it. On account creation, its validated, bounded snapshot is stored with your account until that account is deleted. Withdrawing consent deletes the browser attribution cookie and stops future browser analytics; it does not rewrite the historical signup snapshot. No ad-tracking.
10. Contact
support@whittleos.com