Privacy Policy
Last updated: November 14, 2026
1. Local Browser PDF Processing
Core PDF viewing, editing, merging, splitting, compression, image conversion, and document export run locally inside your web browser using client-side libraries. Source PDF files opened in the core workspace are not uploaded to an application backend.
2. Local Browser Storage
Documents you open in the editor may be stored locally in your browser's IndexedDB storage to power the “My documents” and “Favorites” views on your current device. You can clear browser site data at any time to remove locally cached files.
3. Account Authentication (Downloads & AI Tools)
An authenticated account via our configured Supabase authentication service is required before downloading exported, merged, split, or converted files, accessing the account-gated library views, and invoking server-backed AI tools. Authentication sessions and profile metadata (such as your email address or Google OAuth profile details) are used for account sign-in and session verification.
4. Optional AI Features (Google Gemini API)
The AI Tools workspace provides signed-in users with Summarize (/api/ai/summarize) and document Chat (/api/ai/chat) powered by Google's Gemini API (generativelanguage.googleapis.com).
What is sent: PDF text extraction runs locally in your browser using PDF.js, and the original PDF file is never uploaded. Only the extracted plain text (up to 100,000 characters and 5,000 pages per document; 512 KB total JSON payload) and any chat questions or conversation turns you submit (up to 20 messages, 4,000 characters each) are transmitted to our server endpoints and forwarded to Google's Gemini API solely to generate the requested summary or response.
Retention & logs: PDF Studio's server endpoints request responses without HTTP caching and do not store extracted document text, chat prompts, or AI replies in an application database; active chat messages remain only in your current browser tab. Server logs record operational metadata (request timing, operation name, payload size, and provider-reported token counts) and exclude document text, prompts, responses, and user identifiers. When rate limiting is active, server-side Supabase quota records store request counts, token-unit estimates, and timestamps keyed to your user ID without storing document text or outputs. Google's handling and retention of data sent to the Gemini API are governed by Google's applicable API terms for the configured deployment key rather than an application-enforced retention guarantee.
Your controls: AI processing is optional and requires checking a per-document confirmation box before either Summarize or Chat requests can run. You can remove the loaded document at any time or refrain from submitting confidential, regulated, or sensitive documents you are not authorized to share with Google's Gemini API.
5. Contact
Questions regarding privacy or this deployment can be directed via the Contact page.