Where your data lives.

We try to parse the documents you upload in your browser first. When that works (most desktop browsers), the original files never leave your device. When your browser can't (older iPhones, scanned PDFs, large files), the file is briefly processed on a stateless server route that does not store it. Either way, only the extracted text is kept in your session, in your browser, and sent over HTTPS to the K2 Think V2 API to build your panel and grade your answers.

Our backend is stateless. There is no database, no file storage, no logs of request bodies. Each route handler processes one request and returns one response. The PDF export is the same — your report text is sent to a stateless rendering route, the PDF is returned to your browser, and the request is discarded.

We do collect anonymous, aggregate page-view counts via Vercel Analytics — no cookies, no cross-site tracking, no PII, no link to your documents or interview content. It lets us see roughly how many people are using the product. That is the only thing we measure.

Your session — the brief, the questions, the answers, the report — lives only in your browser through IndexedDB. Clear your site data and it's gone.

The K2 Think V2 API is operated by MBZUAI / their hosting provider. Whatever they retain in their API logs is governed by their terms. We send only document text to them; we do not attach identifiers beyond the text itself.