We process only the data needed to provide the product's features. This may include account identifiers needed for authentication, PDF URLs and document identifiers, local extension state such as a per-PDF viewer preference, user-selected text from PDFs, notes, annotations, collections, saved document metadata, AI prompts, follow-up questions, AI responses, citation metadata, citation context, reference text, billing identifiers for paid plans, and basic usage or error information needed to keep the service reliable.
Usage and reliability data may include event names, timestamps, browser and device metadata, extension or app version, page or feature context inside Wavely, error messages, stack traces, and account identifiers where needed to debug signed-in issues. We do not intentionally send full general browsing history or unrelated non-PDF page content to analytics or error-monitoring tools.
Where diagnostics or monitoring payloads include viewer URLs, we aim to redact document-routing query parameters such as the current PDF file URL before sending them to monitoring tools.