Write on PDF

Privacy policy

Updated November 2, 2026. This policy covers the Write on PDF website, its online PDF editor and browser extension.

PDF editing and browser storage

PDF rendering, annotations, drawn signatures, image conversion and PDF export run on your device. Settings, annotation drafts, recognized text, text-editing drafts, comments, styles and saved signatures may be stored in your browser. Website data uses IndexedDB and is separate from extension storage. A PDF selected on the homepage is held temporarily in browser storage to pass it to the editor, then removed from that temporary store. Image pages inserted into a document may also be retained locally as part of its draft. The original open document stays in memory. Passwords are used in memory to open and export protected documents.

Clear the site’s browser data to remove website drafts and signatures. Removing the extension removes its extension storage. Downloads and files saved to Google Drive are not removed by either action.

Digital signatures and optional service seals

Signing with your own certificate processes the certificate file, private key and password locally. They are not uploaded or saved by the editor. The public certificate is embedded in the downloaded PDF.

Choosing the Write on PDF service seal sends a SHA-256 fingerprint of the PDF’s signed byte ranges, a random record ID and the public service certificate fingerprint to our Cloudflare Worker. PDF contents, file names, document passwords and personal certificates are not sent. The shared service private key stays on the server. Cloudflare D1 retains the record ID, signed-content fingerprint, certificate fingerprint, server registration time and public digital signature so that the record can be checked later. These records are retained for verification and are readable by anyone who has the random record ID; there is no public record directory.

When you open a signed PDF, signature integrity is checked locally. Choosing Check signed version in Signature details sends its record ID to our server and compares the returned fingerprint on your device. It does not upload the PDF. A daily keyed IP hash and attempt count are used for service seal limits; old quota buckets are periodically removed. The application does not store raw IP addresses in seal records or quota tables. Cloudflare may process technical request metadata under its own policies.

Optional Google Drive connection

The website loads Google’s Picker library when the editor opens so the file chooser is ready. Loading the library does not sign you in or grant access to your files.

Drive is optional. When you choose a Drive action that needs sign-in, Google opens in a separate browser window to sign you in and request file access. After that window closes, the Google file chooser opens inside the editor. We request access to the Drive files you choose or create with Write on PDF and permission to connect this app to Drive. We do not request permission to list or read all your Drive files, Gmail or contacts. Google provides your basic profile and email to create or restore your Write on PDF account in Firebase Authentication. Ordinary local editing does not create an account.

The browser reads the selected PDF’s contents, name, size and relevant file permissions directly from Google. Choosing Save copy to Drive uploads the exported PDF and its name directly to Google as a new file in My Drive. We do not change the source file or its sharing settings. These file transfers do not pass through our server. Google’s scripts and services process sign-in, file selection and Drive requests under their own policies. The embedded Google file chooser may request Google cookie access, including in Safari.

Firebase Authentication performs the Google sign-in and remembers your account in this browser. Our server checks that the Google Drive credential belongs to that Firebase account before connecting it. Google access tokens are used in tab memory and may also be held encrypted on our server while establishing or restoring the connection. Only unexpired tokens are used. Existing encrypted Google refresh tokens remain linked to their Firebase user IDs in Cloudflare D1 and are never returned to the browser. The current Firebase sign-in and file chooser do not create a new Google refresh token; if an access credential expires and no saved refresh credential is available, a later Drive action asks you to reconnect. Firebase maintains its own separate authentication session; its session credential is not used as a Drive credential. Temporary sign-in request identifiers stay in the sign-in window’s session storage. Selected file identifiers are briefly held in browser storage to return the result to the editor, then removed; abandoned handoffs expire after ten minutes. Signing out clears the local Firebase session. Disconnect deletes the saved Drive credential and requests Google revocation; you can also remove the app in Google Account connections. PDF drafts may remain in local browser storage until you clear them.

Information received from Google Workspace APIs is used only to provide the document actions you request. We do not sell it, use it for advertising or use it to train AI models. Write on PDF’s use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including its Limited Use requirements.

Optional cloud text recognition

Choosing Edit text asks for confirmation before uploading the selected page through our Cloudflare API to Claude on Google Vertex AI. This also applies to a PDF opened from Drive. Added editor notes and signatures, source annotations and form widgets, other pages and passwords are excluded. PDF reconstruction and export happen locally. To keep a Drive document entirely local during editing, do not enable this cloud action.

The scan API does not store or log PDF contents or recognized text. Cloudflare and Google process requests under their applicable service terms. To enforce the scan quota, Cloudflare D1 stores a daily keyed IP hash, day and attempt count. Raw IP addresses are not stored in that table; providers may process technical request metadata. Old quota buckets are removed by a periodic cleanup job.

Website requests and optional feedback

The website and extension include no analytics or advertising trackers. Fonts are bundled locally. Opening a PDF from another website requests the file from that website and may use the existing sign-in session. Our hosting provider may process technical request information to serve and protect the site.

Optional uninstall feedback stores the selected reason, optional comment, submission time and a random response ID in our Cloudflare-hosted database. PDF contents and browsing history are not attached. You can close the feedback page without answering. Do not include sensitive information in a comment.

Questions

For privacy or support questions, email wpdf@fonable.io. Do not send confidential documents or Google access tokens.