Privacy

Your file never leaves your browser. When you select a .pptx, it is parsed entirely in your browser, locally on your device. The file itself is never uploaded to Sharp or to anyone else.

Extracted text is sent to a commercial AI API. To generate the analysis, the extracted slide titles, body text, speaker notes, and structural counts (fonts, sizes, colors, word counts) are sent to a commercial AI API. That is the only content that leaves your device, and it is not stored by Sharp.

What the AI provider does with it. The analysis runs on Anthropic's Claude API. Under Anthropic's commercial terms, API data is not used to train their models without express permission, and prompt and response content is not retained by default — with limited exceptions for safety and legal compliance. See Anthropic's data retention documentation for the current policy. Sharp has a standard API relationship with Anthropic; your text is processed to produce your report and nothing more.

Sharp stores nothing. Sharp has no database. We do not save your deck content, your reports, or any record of what you analyzed. When you close the tab, everything is gone.

No content analytics. We do not log, mine, or train on your deck content.

A practical recommendation. Because the extracted text passes through a commercial AI API, we recommend not uploading confidential or client material. Sharp is ideal for drafts, practice decks, and sanitized content — strip client names and sensitive numbers before grading.

Payments. The $7 unlock is handled by Stripe on Stripe's hosted checkout page. We never see your card details. Stripe's own privacy policy governs the payment itself.

Questions: support@sharpdecks.com