How your files are handled

• Only the account that owns a file can download it.
• Downloading needs a signed-in account. Files made before signing in are deleted after 1 hour unless you sign in; account files stay until you delete them.
• Files are stored encrypted in Amazon S3 in Mumbai, India, and processed on our own servers there.
• No outside conversion services, no advertising and no AI training on your files. Google Analytics counts page visits but never receives your files.

Files travel over HTTPS from your browser to our servers. Before anything is processed, each file's type is checked against both its name and its contents, and its size and page count are checked against the limits (50 MB and 1,000 pages).

Every download needs a signed-in account, and only the account that owns a file can download it. Files you make before signing in are linked to your browser by a private cookie and move into your account when you sign in there. Knowing a file's ID isn't enough to download it.

A share link is the only way to give someone else access. Links are long random codes that we store only as a hash. Each works for up to 7 days, never longer than the file is kept, shows how many times it's been downloaded and can be revoked at any time. Anyone who has the link and signs in can download the file until then.

Files are kept in a private Amazon S3 bucket in AWS's Mumbai region. Public access is blocked, every file is encrypted at rest with AES-256 and connections must use HTTPS. Our processing servers reach the bucket through a restricted access role, with no access keys stored on them. File records and account details are in our database, hosted by Neon on AWS in Singapore.

• Every job runs in its own process on our servers, with a memory limit and a 5-minute time limit, so one file can't affect another.
• PDF work uses open-source engines (pdfcpu and PDFium), Office files are converted with LibreOffice, web pages are rendered with Chromium and text is recognised with Tesseract. Nothing is sent to another company's conversion service.
• Uploaded HTML files are rendered with no network access. When you convert a web address, our server can reach only public internet addresses, never private networks.

• Accounts sign in with Google, so PocketPDF doesn't hold a password for them. Accounts created earlier with a password keep it only as an Argon2id hash. Session IDs are stored only as hashes.
• The passwords you give Protect PDF and Unlock PDF are passed straight to the job, never shown again and removed from our records as soon as the job finishes.

• Before signing in: files, results and their activity records are deleted automatically after 1 hour, unless you sign in on the same browser.
• In an account: delete a file, several files or everything from Documents. The stored copy is removed straight away; storage keeps no older versions.
• Deleting your account removes all of its files, share links, activity and usage history.
• Activity history in an account covers the last 30 days.

• No online service can be perfectly secure. For very sensitive documents, consider whether you need to upload them at all, and redact what you don't need to share.
• Uploads aren't scanned for viruses, so open shared files only from people you trust.
• Redaction, OCR and conversions are automatic; check the result before you send it on.

If you find a vulnerability, email hi@pocketpdf.in with "Security" in the subject and the steps to reproduce it. Please don't access other people's data or disrupt the service while testing, and give us reasonable time to fix the issue before you make it public. We'll acknowledge every report.