Log Analyzer

Make sense of a log file without grep. Drop one or more logs, or paste the text, and every line is colored and counted by level — errors, warnings, info, debug — with stack traces kept together with the error that caused them. Filter by level, text or regex, see which errors repeat most once IDs and numbers are masked, and spot the busiest IPs and status codes. Large files stay fast. It's free, runs in your browser, and your logs never leave your device.

Level words recognized

LevelMatches
fatalFATAL, CRITICAL, EMERG, ALERT, PANIC
errorERROR, ERR, SEVERE, FAILED
warnWARN, WARNING
infoINFO, NOTICE
debugDEBUG, DBG
traceTRACE, VERBOSE

When to use it

  1. Find the error behind an outage in a long application log.
  2. See which warning fills the log most often.
  3. Pull every 500 response and its client IP from an access log.
  4. Share just the relevant lines by exporting a filtered view.

How to Use

  1. Drop log files or paste log text.
  2. Toggle levels and search, optionally with a regex.
  3. Review the most repeated errors and the IP and status tallies.
  4. Export the filtered lines if you need to share them.

Features

  1. Level detection and counts
  2. Stack traces grouped
  3. Text and regex filtering
  4. Repeated-message grouping
  5. IP and status code tallies
  6. Fast on large files
  7. Multiple files
  8. 100% client-side

FAQs

Which log formats does it understand?

Plain text logs with level words, JSON logs with a level field, and Apache/Nginx access logs. Timestamps in ISO 8601, access-log and syslog formats are recognized.

How are stack traces handled?

Indented lines and lines starting with at, Caused by or Traceback are attached to the entry above, so they're kept with their error.

What does 'most repeated' mean?

Messages are compared with numbers, IDs, IPs and times masked, so the same error with different values counts as one.

Can it handle big files?

Yes. Only the visible lines are drawn, so scrolling stays smooth on very long logs.

Are my logs uploaded?

No. Everything is processed in your browser.

Technical Details

Each line is classified by level from the words logs actually use — FATAL, CRITICAL, ERROR, SEVERE, WARN, INFO, NOTICE, DEBUG, TRACE — and from structured JSON logs' level field, including numeric pino/bunyan levels. Lines that begin with whitespace, at, Caused by, Traceback or similar are treated as continuations and attached to the entry above, so a Java, Node or Python stack trace stays with the error that produced it and is filtered with it. ISO 8601, Apache/Nginx and syslog timestamps are recognized. To find what's really going wrong, each warning and error is reduced to a shape by masking numbers, UUIDs, hex IDs, IP addresses and timestamps, and identical shapes are counted, so a thousand timeouts for different request IDs show as one message seen a thousand times. IP addresses and HTTP status codes are tallied too. Filtering by level, text or regular expression is instant, and the viewer only renders the lines on screen, so logs of hundreds of thousands of lines scroll smoothly. Several files can be combined, and the filtered view can be exported.

Developer Tools

Developer Tools

  • .env File Validator
  • HTTP Headers Parser
  • Cookie Parser
  • Dockerfile Generator
  • Nginx Config Generator
  • README Generator
  • Dockerfile Linter
  • SSH Config Generator