Privacy boundary
Ordinary PdfEditorOnlineFree document workflows are implemented to process selected files, decoded pages, extracted text, previews, and generated outputs in browser code rather than a PdfEditorOnlineFree conversion endpoint. Repository evidence and initial-load browser audits support that architecture; each tool's narrower disclosure and operation-level evidence define the exact boundary. This does not mean every tool makes no network request.
Most file bytes, rendered pages, intermediate objects, previews, and outputs live in browser memory for the current tab or operation. Replacing a file, clearing a workspace, closing the tab, or browser cleanup can discard them.
Some tools save opt-in settings, templates, verification manifests, model caches, or metadata. Business and peer tools may use IndexedDB; model and runtime reuse may use browser Cache Storage. Clearing site data removes these records.
PDF engines, OCR data, fonts, workers, and optional office runtimes are application code, not the selected document. Same-origin runtime and worker files can be requested on first use and cached by the browser.
AI features disclose size, revision, license, storage, and integrity details before setup. After explicit consent, pinned external model hosts may receive ordinary asset GET requests. Document bytes and extracted document text are not model-download payloads.
Live peer-to-peer flows cross the network between participating browsers. File or collaboration data leaves the originating device for the chosen peer, even when it does not pass through a PdfEditorOnlineFree content server.
Google Analytics measures page visits, referrers, timings, and tool outcomes such as file type, size bucket, and duration. Session replay and source-content analytics stay disabled, file names and file contents are never sent, and you can reject measurement below.
Measurement
After your first successful output download, an optional “How did you hear about us?” prompt may appear only when analytics consent is granted. It uses fixed choices, has no free-text field, and collects no document data or other personally identifiable information. Your browser stores only whether a first download occurred and whether the prompt was shown or resolved so it does not repeat. Answers are a supplementary signal from a self-selected, analytics-consenting subset and are not representative of all visitors or a replacement for observed attribution.
What to review per tool
A simple PDF operation, a 300 DPI OCR run, a large local AI model, a persistent point-of-sale database, and a WebRTC file transfer do not have the same storage or network profile. Read the visible privacy, asset, memory, output, and limitations sections on the tool you intend to use.
Browser extensions, device management software, operating-system services, the network itself, and the browser vendor remain outside PdfEditorOnlineFree's application boundary.