Website privacy.

The site uses no analytics scripts, advertising pixels, or tracking cookies. It is hosted on Vercel, which processes standard server logs, including IP address, user agent, and request timestamps, to serve pages and to run security and reliability checks. Fonts are served from tryraptor.app, so no font network sees your visit. The home page animation runs in your browser and reports nothing back.

On desktop browsers, the home page ends in a short conversation with a site assistant; on phones the page shows an email address instead. The assistant answers questions about Raptor and declines other topics. Each time you send a message, the conversation so far goes to Raptor's server and from there to a model provider, OpenRouter, and the model it routes to, which generates the reply. OpenRouter and the model provider process the text under OpenRouter's terms. Raptor does not use conversations to train models.

For each reply, Raptor's server logs a request id, the model name, the number of turns, the response time, token counts, and whether the reply was withheld by Raptor's own filter. It does not log what you typed or what the model answered. To cap requests per hour, the server keeps a hash of your IP address in memory, keyed with a value that exists only while the server runs; the hash is never written to storage. Raptor stores nothing from the conversation unless you send the card at the end of it. Do not enter passwords, payment details, or other sensitive information in the conversation.

When you send the card, Raptor stores your name, email, the conversation up to that point, the page you sent it from, the time you sent it, a salted hash of your IP address, and whether a hidden anti-spam field was filled. Submissions are held in Raptor's database on Supabase and capped per address. Andy reads every submission and replies from his own account; nothing is sent automatically.

A submission is erased on request: the personal fields are blanked and only the arrival date and source page are kept so counts stay true. Email links on the site open your email client, and your provider and Raptor's, Google Workspace, process anything you send that way.

The client portal is by invitation and is not linked from this site. Raptor creates every account directly and there is no self-serve signup. Sign-in uses your Google account through Supabase Auth, which shares your name, email address, and profile photo with Raptor to identify your account and nothing else. Supabase stores your account email and sign-in records. Raptor stores which client workspace you belong to and your role in it, and shows you the status of the work Raptor is doing for your organization.

The portal sets session cookies so you stay signed in, and signing out ends the session. It uses no analytics or tracking cookies and collects no payment details; invoicing happens outside the portal.

To ask what information Raptor holds about you, or to request correction or deletion, email andy@tryraptor.app. Raptor retains records it needs for security, legal, or accounting reasons. This notice is updated as the site changes, and the effective date above reflects the current version.