Log and Privacy Workbench

This first release auto-detects JSON Lines (one JSON object per line), a JSON array of log objects, and plaintext server logs split by line. For each entry it makes a best-effort attempt to extract a timestamp, level, service/logger name, request/correlation ID, and message.

Continuation lines that look like Java/Node stack frames or Python traceback lines are attached to the previous entry instead of treated as brand-new entries. Repeated traces are then grouped by exception class name and the first one or two frame lines, which works well for identical recurring failures but is still heuristic.

Not yet. Repeated-message grouping and compare mode both use exact message text in this release, which keeps the behavior predictable and fast but means small wording changes create separate groups.

No. Everything runs locally in the browser. Raw log text is not sent to a server, not stored in the page URL, and not persisted in local storage by this tool.

Before exporting, the tool runs the current output through the existing secret detector and optionally replaces matched values with stable placeholders such as [REDACTED-JWT-1]. That preserves repetition patterns without exposing the original credential or token.

To keep the tab responsive, input is capped at roughly 5 MB or 50,000 lines. If you paste more than that, the tool shows a truncation notice and parses only the kept portion instead of trying to render everything.