Privacy Policy

PriorLens ("we", "us", "our") is a prior-art research tool that helps inventors and researchers search US patents using USPTO-provided patent data services and assess overlap with their own claims. Contact: privacy@priorlens.com.

When you sign in, authentication is handled by Neon Auth (Neon's identity service). Neon Auth verifies your credentials and issues a signed session token that includes your unique user identifier, email address, and display name. Our servers read those fields from that token to show your name in the UI and to route saved scans back to you; your account/identity record itself is created, stored, and managed by Neon, not on our own servers.

When you save a scan (the "Save scan" action on the claim chart), we store the search query you ran, the parsed claims, the ranked references returned by USPTO, the AI-generated rationales, and any annotations or notes you've added. These rows are owned by your user identifier and visible only to you, unless you create a share link (see §4).

For authentication, your browser holds a short-lived Neon Auth session token and attaches it as a Bearer token on requests to our API. Our servers verify that token's signature on each request; we do not store session rows or refresh tokens ourselves.

For abuse prevention we read your IP address from request headers and use it for per-IP rate limiting (uploads, AI queries, USPTO calls). IPs are not stored long-term, but they do appear in short-lived request logs.

For accountability we record a small audit log of sensitive actions (saving a scan, deleting a scan, creating or revoking a share link, deleting your account, and visits to shared scans). Each row includes a timestamp, action, related identifier, and IP.

We use the data above only to operate the service: render your account, run searches, save scans you choose to save, prevent abuse, and respond to support requests. We do not sell or rent personal data. We do not use your queries or saved scans to train AI models.

Patent data services (epo.org): your search query is forwarded to the European Patent Office's Open Patent Services (EPO OPS) so it can return matching patents, worldwide. USPTO's legacy PatentsView search API was retired on March 20, 2026; EPO OPS is the current, live search backend. Subject to EPO's applicable data-handling terms.

Anthropic: when AI query expansion, per-claim rationale, prior-art report generation, or the provisional-application drafting assistant is used, the relevant invention/claim text is sent to Anthropic's Claude API to generate that result. We only send this text to Anthropic once our account is confirmed to be on a paid tier with no-training terms. We do not include your name or email in those requests.

OpenAI: when semantic search or claim-chart re-ranking is used, your claim text and the candidate patent text are sent to OpenAI's embeddings API, under the same paid-tier requirement. We do not include your name or email in those requests.

Gab AI (fallback, used only if Anthropic is unavailable): your query text and claim text are sent to Gab AI's chat completions API under the same paid-tier requirement. Disabled by default.

Before any of the text above is sent to Anthropic, OpenAI, or Gab AI, our server automatically strips email addresses, phone numbers, and street addresses it detects — an extra precaution on top of the paid-tier requirement, not a substitute for it. This detection is pattern-based and not perfect; please avoid including your own contact details in a query or description if you'd rather they never leave your device at all.

Neon: our authentication and database provider. Neon Auth handles sign-in and holds your identity record; Neon Postgres stores our application data (saved scans, annotations, share links, usage counters). Subject to Neon's privacy policy.

Fly.io: our hosting provider. They run our application servers. Subject to Fly.io's privacy policy.

Stripe: our payment processor for the paid Invention Prior-Art Report. Subject to Stripe's privacy policy.

Share links: when you click "Share link" on a saved scan and confirm an expiration, the server issues an unguessable public URL that anyone can open without signing in. That URL exposes the saved scan in full (query, parsed claims, ranked references, AI rationales, and your notes). You can revoke a share link from Settings at any time.

Saved scans, annotations, share links, and your account row persist until you delete them. You can:

• Delete an individual saved scan from the recent-scans list on Home.
• Revoke a share link from Settings → Data & Privacy.
• Clear local browser data (counters, cached results, preferences) from Settings → Clear Local Data.
• Delete your account data (cascades scans, annotations, share links, and usage records) from Settings → Delete My Account. This removes the data we store on our servers; your Neon Auth sign-in identity is managed separately by Neon.
• Export everything we store about you as JSON from Settings → Export My Account Data.

Automatic invention-text purge: independent of the manual deletion above, we automatically delete saved scans, and clear the stored description and generated report on Invention Prior-Art Report orders, 90 days after creation (report orders: 90 days after the order reaches a final delivered/failed state). This window is configurable on our servers; 90 days is the default and what's running in production as of this policy's Effective Date.

Server logs and audit records age out within 30 days of creation unless we are legally required to retain them longer (for example, in response to an abuse complaint or subpoena).

Production traffic runs over HTTPS with HSTS. We apply strict security headers (Content-Security-Policy, X-Content-Type-Options, Referrer-Policy, and others) and use a double-submit CSRF token on state-changing requests. Authentication uses short-lived Neon Auth session tokens verified on every request; we do not store session rows or refresh tokens on our servers. Third-party API keys (Anthropic, OpenAI, Gab AI) live only in server environment variables and never reach the browser.

Patent searches are performed with a server-held USPTO API key; you do not need to provide, and we do not store, any API key of your own in the browser.

No system is perfectly secure. If you believe you've found a vulnerability, please email security@priorlens.com.

PriorLens is not directed at children under 13 and we do not knowingly collect personal information from children. If you believe a child has used the service, please contact privacy@priorlens.com so we can delete the relevant data.

You can access, export, and delete your data using the buttons in Settings (see §5). You may also email privacy@priorlens.com with a request and we will respond within a reasonable time. Depending on where you live, you may have additional rights (for example, the right to lodge a complaint with a supervisory authority).

PriorLens is operated from the United States. By using the service from outside the US, you understand that your data will be transferred to and processed in the US.

We may update this Privacy Policy. The Effective Date above will change when we do. Material changes will be announced in-product.

Questions? Email privacy@priorlens.com.