Logs & observability tools
Paste or open a log file, click Format, then filter by level, search, and switch how timestamps are shown — all on the already-parsed rows.
Runs locally in your browser. Nothing you paste or open is uploaded.
Format a log to see rows here.
A single grep gets you a handful of matching lines, but it stops being enough the moment a support ticket contains a mix of structured JSON events, a stray logfmt line from a middleware, and a Kubernetes container's stdout all pasted together. Working out which lines are errors, what a garbled Unix timestamp actually means, and which requests match a given user ID by eye is slow and error-prone.
This log viewer is for the moment you have that mixed paste in your clipboard and need a readable table instead of a wall of text: it detects the shape of each line on its own, extracts a level and a message where it can, and lets you narrow down to what matters without re-running anything. It is not a replacement for a real log aggregator watching a live pipeline — it works on a static block of text you paste or open from a local file, once, in the browser tab you are already in. Some people call this kind of tool a log viewer, others a log formatter; either name points at the same job: turning pasted text into a readable table.
Built around one mixed paste, not one format at a time.
JSON events, logfmt key=value pairs, a Kubernetes container's CRI-formatted stdout, a Docker json-file driver line, and plain text can all sit in the same paste — Auto-detect classifies each line on its own instead of forcing one format for the whole block, and a line that matches nothing still shows up as plain text instead of being dropped.
Formatting a fresh paste is a deliberate click (or Ctrl/Cmd+Enter), but once the table exists, checking or unchecking a log level, typing a search term, and switching between raw, local, UTC and relative timestamps all recompute immediately on the already-formatted rows — no need to click Format again for any of that.
Three steps, and what each one actually does under the hood.
Paste log text directly, or click Open file to read a local .log/.txt/.json file with the browser's file reader — the file is never uploaded anywhere. Format preset defaults to Auto-detect, so most of the time you do not need to pick a shape yourself, but if you already know every line is, say, logfmt, forcing that preset skips the detection step entirely.
Formatting does not run automatically while you type — re-detecting the shape of every line on each keystroke would be wasted work for a paste that is still being edited. Click Format once (or use the shortcut) and every non-blank line becomes a row: click a row's Fields button to see exactly what was extracted, including the raw fields from a JSON object or logfmt pair.
Level checkboxes start fully checked, so nothing is hidden until you uncheck one — and a line the tool could not confidently classify with a level always stays visible regardless of which boxes are checked, since hiding unclassifiable content would be a guess, not a filter. Search can switch to regex, and the timestamp column can switch between raw, local time, UTC, and relative — none of which touch the text you pasted. Copy or download exactly the currently filtered rows as plain text or structured JSON with a parsed epoch value.
Detection, filtering and export built for a paste you did not write the parser for.
JSON, logfmt, Kubernetes CRI, Docker json-file, and plain text are each detected independently for every line, so one mixed paste from several sources still produces one consistent table instead of forcing you to split it first.
Narrow the table to specific severities and a search term at the same time, with an optional regex toggle, all recomputed on the already-formatted rows without re-running detection.
Switch the timestamp column between the raw text you pasted, your local time, UTC, or a relative phrase like "3 minutes ago" — it only changes what that column shows, never the original text sitting in the input box.
Expand any row to see every field the detector pulled out — a JSON object's keys, a logfmt pair, or a Kubernetes stream and tag — and export the currently filtered rows as plain text or as structured JSON that includes a parsed epoch timestamp.
Real limits, not a marketing list.
Answers to the questions people ask before pasting logs into an online tool.
Yes. Paste or open your log, click Format, and the table appears immediately in this log viewer — no account, install, or upload required. It runs as JavaScript in your current browser tab.
Nothing you paste or open is sent anywhere. Formatting, filtering, searching, and exporting all happen locally in your browser tab; there is no server-side upload, storage, or logging of the text, which matters because real logs often contain user IDs, IPs, and other sensitive details.
logfmt is a flat key=value format on one line (level=info msg="done" duration=12ms), designed to be both human-readable and easy to grep. JSON log lines are one JSON object per line, more structured and easier to parse programmatically but harder to scan by eye. This log viewer detects and extracts fields from both automatically, so you do not have to pick a parser ahead of time.
Once you format a log, switch the Timestamp dropdown to Local time, UTC, or Relative and every row's detected timestamp re-renders in that format — the raw text in the input is never changed. For converting one timestamp on its own outside a log, use the site's dedicated Unix timestamp converter instead.
No. "Kubernetes container log" and "Docker json-file log" are format presets that tell this log viewer which line shape to expect when you paste text you already copied out — they are not a live connection, require no credentials, and the tool makes no network requests at all. For live tailing of a running cluster or container, you need a different, connected tool.
Type a term into the search box and the table narrows instantly to lines containing it, without re-running Format. Turn on the regex toggle next to it to search with a regular expression instead of a plain substring, and combine it with the level checkboxes to narrow by both at once.
Found a real log line this tool gets wrong?
Paste a mixed block of JSON, logfmt, Kubernetes, Docker, or plain-text log lines and get a filterable, exportable table in seconds — no upload, no account.
Bookmark this free log viewer for the next time a support ticket or a terminal dump lands in your clipboard as a wall of unreadable text.
All formatting happens locally in your browser.