All Tools View Categories About Contact Privacy

Log Digest Generator

Turn raw log lines into a daily or weekly digest — aggregated counts by category, level and time.

Runs entirely in your browser — logs never leave this page.

About Log Digest Generator

Log files are the first place a team looks when something breaks — and the last place they want to spend an afternoon. A thousand lines later you still do not know the shape of the problem: which service logged the most errors, whether failures clustered around a deploy window, or whether today is genuinely worse than yesterday. Reading raw log lines by eye is exactly how real incidents get missed.

Log Digest Generator turns that wall of text into a readable daily or weekly digest in one paste. It parses each line, detects the timestamp, severity level and service or category, then aggregates counts by category, by level and over time — so you see how often and where instead of re-reading everything. Bracket formats, plain timestamp LEVEL category lines and JSON logs are all detected automatically, with no format configuration required.

Because every system formats logs differently, the options do the heavy lifting: pick a digest period (daily, weekly or monthly), limit which levels are counted, narrow the window to a date range, and drop in a custom regex when categories live in an unusual place. The output renders as summary cards, a bar chart of the top categories and a level breakdown, with the full digest exportable as plain text, Markdown, JSON or CSV for your standup, alerting or on-call report.

Everything runs in the browser — paste real logs with confidence; nothing is uploaded.

Features

  • Auto-detecting parser: bracket, plain and JSON log lines recognized with no setup.
  • Timestamp, level and category extraction: the digest is built from what each line really says.
  • Daily, weekly or monthly period: the digest matches your review cadence.
  • Counts by category: top services and areas with error, warning and share totals.
  • Level breakdown: fatal, error, warn, info, debug and trace in one table.
  • Time trend: busiest hour for daily digests, per-day counts for weekly and monthly.
  • Level and date filters: exclude noise and window to a specific incident.
  • Custom category regex: extract categories from unusual log layouts.
  • Multi-format export: plain text, Markdown, JSON or CSV — copy or download.
  • Sample logs: load a realistic dataset in one click to try every feature.

How to Use

  1. Load sample or paste logs. The Load sample logs button fills a realistic two-day dataset spanning several categories and levels so every feature is visible immediately — or paste your own logs.
  2. Choose the period. Daily, weekly or monthly. The trend and busiest-period labels adjust automatically.
  3. Tune the filters. Deselect levels you want to ignore and narrow the date range to an incident window.
  4. Add a regex only if needed. Default detection handles bracket, plain and JSON formats; a custom regex with a capture group overrides category extraction.
  5. Read the digest. Summary cards, category bars and the level table appear instantly, and the full report is ready to copy or download in your chosen format.

Examples

Example 1 — Deploy watch. A release went out at 14:00. Paste the day’s logs, set the period to Daily, and the busiest-hour card plus the api category’s error column immediately reveal whether the new code correlates with the spike.

Example 2 — Weekly on-call report. Every Friday the on-call engineer pastes the week’s logs, sets Weekly and exports Markdown. The per-day trend shows which days were noisy and the category table shows which service generated the most alerts.

Example 3 — JSON pipelines. Services that log structured JSON lines parse cleanly. The digest extracts ts, level, category/logger and message keys, so counts reflect real service boundaries instead of string matching.

Example 4 — Sifting out noise. Debug and trace lines dominate raw volume. Deselect those levels and the digest narrows to warn and above, surfacing the signal an incident response actually needs.

Example 5 — Custom layouts. Legacy apps log categories inside their own delimiters. One regex like \[([a-z_-]+)\] captures the category and every count updates.

Benefits

  • See the shape of your logs in seconds, not by scanning.
  • Errors and warnings by category make the affected service obvious.
  • Daily and weekly views catch slow-burn trends, not just spikes.
  • Filters cut the noise before the digest is built.
  • Text, Markdown, JSON and CSV fit your reporting stack.
  • Fully private: logs are parsed locally and never uploaded.

Frequently Asked Questions

What log formats does the generator support?
Bracket formats like [2026-08-14 09:00:01] [ERROR] [api] message, plain lines like 2026-08-14 09:00:01 ERROR api message, and JSON lines with ts, level and category/logger/message keys. Auto-detect tries JSON first, then the common text formats.
How are categories identified?
By default the first bracketed token that is not a timestamp or a level is treated as the category, so [api] and [db] become groups. A custom regex overrides this — anything in capture group 1 becomes the category.
Does every log line need a timestamp?
No. Lines without a timestamp are still aggregated by category and level; only the time-trend and busiest-period cards are omitted for them.
What does the level filter do?
It controls which severity levels are aggregated. Deselecting debug and trace removes their volume so warn and error counts stand out in the digest.
How is the share percentage calculated?
Each category’s count divided by the total of included entries after all filters, so percentages always sum to 100 across the displayed set.
Can I narrow the digest to one incident?
Yes. Set the From and To date inputs and only lines whose timestamps fall inside the window are counted — ideal for isolating a deploy window or an outage.
Is my log data uploaded?
No. Parsing and aggregation run entirely in your browser; nothing is uploaded, stored or logged. Paste real production logs with confidence.