JSON Log Viewer & Analyzer

JSON Log Viewer & Analyzer turns structured logs into a searchable event list. It detects the timestamp, level, message and service fields, normalizes levels such as pino numbers or syslog words, and shows every event with its nested JSON intact, so you can find the failing request in a production log without writing jq or grep pipelines.

JSON arrays, JSONL / NDJSON (one object per line), Docker json-file logs (the inner log field is parsed when it is JSON itself) and mixed files. Lines that are not valid JSON, such as stack traces or plain text, are kept as text events with a parse-error marker instead of being dropped, and JSON after a text prefix (for example a timestamp and container name) is still recognized.

Combine a full-text or regex search, level checkboxes, a time range and field conditions such as status >= 500, service = payment or -trace_id = abc123. Conditions support =, !=, ~ (contains), !~ and numeric comparisons, with dotted paths for nested fields. Selecting an event lists its trace, request, correlation, service and user identifiers with one-click Filter, Exclude and Copy actions.

Files are read in 1 MB chunks by a background worker, so the page stays responsive while multi-megabyte logs load. Line-based logs up to 200 MB and 1,000,000 lines are supported; JSON arrays up to 50 MB, because an array must be parsed in one piece. Only the visible rows are rendered, and filtering runs in the worker with a time limit for slow regular expressions.

Logs are parsed, filtered and exported entirely in your browser. Nothing is uploaded, and log contents, filters and field mappings are never stored; only the row density and time zone preferences are remembered on this device. Exported JSONL contains exactly the events that match your filters.