One platform for collecting, searching, understanding, and acting on your logs, plus browser automations that catch a broken sign-in before your customers do, and the two features nobody else has: git regression detection and zip-the-diff patterns.
When an error carries a stack trace, ZipLogger goes further than any log tool you've used:
it parses every file:line frame, runs git blame against the repositories you've
connected, and produces a ranked list of suspect commits, weighted by how recently they touched
the exact lines that are now failing.
With AI enabled, each regression case gets a written root-cause analysis and a suggested fix, ready to paste into your incident channel.
# Regression case #42: checkout NullReference
Suspects (ranked):
1. a3f92c1 refactor currency handling
dana@acme · 2h before first error · score 0.91
2. 88d01ce bump payment SDK to 4.2
lee@acme · 2d before · score 0.34
AI: `Currency` can be null when the cart was
created before the deploy. Guard in
PaymentProcessor.Charge() or backfill carts.Logs tell you a request failed. They do not tell you the sign-in button stopped working, or that checkout breaks on step four. A browser automation does: record a real journey once, and ZipLogger replays it in a real browser on a schedule, from every minute to once a day.
Recording happens in a browser that opens right in your tab. Sign in, search, add to basket, check out; every click and keystroke becomes a replayable step. No extension, no download, no test framework to learn. Passwords are captured as a reference, never as text.
// Checkout — sign in and add to basket
every 5 minutes · 11 steps
✓ Open https://shop.example.com
✓ Fill Email · Fill Password (secret)
✓ Click Sign in
✓ Search "espresso"
✗ Click Add to basket
could not find the element
(tried role=button name="Add to basket")
Your server: shop-api POST /cart
NullReferenceException — Cart.AddItem
Likely cause: 2 commits landed since this
journey last passed; a4f21c9 renamed the
basket button. Screenshot attached.Two AI features, both practical. Summarize errors turns any time window into a short root-cause narrative: what broke, where, how often, since when. Ask converts plain English into real queries: “warnings from the API since last night” just works.
"why are checkout errors spiking?"
→ severity:error source:checkout, last 2h
Summary: 214 NullReference errors from
PaymentProcessor.Charge(), starting 14:10,
right after release 2.4.1. Affects ~3% of
checkouts. One pattern, one suspect commit.ZipLogger's pattern engine groups similar messages into templates in real time. Instead of scrolling 400,000 lines, you read ten patterns, each showing its count, trend, and exactly which parts vary between occurrences. It's compression for your attention.
142,801 Order <*> created — <*> items, $<*>
98,112 GET /api/<*> completed in <*> ms
214 Payment failed: <*> — Charge() ⚠
12 Retry <*>/3 for order #<*>Every service, container, and job logs to one place, and search stays fast as your volume grows, with retention managed automatically by your plan. Full-text, field filters, time ranges, histograms: everything composes.
severity:error source:checkout release:2.4.1
→ 214 events in 38 ms
fields.orderId:83112
→ the full story of one order, all servicesCompose timeseries, stat, and top-values panels from any query. Watch error rates, top sources, and release health on one screen, saved per team, no extra product to buy.
Above/below-threshold, no-data, service-down, API latency-increase, and API traffic-anomaly rules, evaluated continuously. Delivery by email, Slack-compatible webhook, SMS or voice call, or a custom sandboxed script with encrypted secrets, plus a full history of every firing. Phone and email notifications work with zero setup (20 calls, 50 SMS, and 100 emails included every month, then prepaid credits), or connect your own Twilio for unlimited sending.
Point any OpenTelemetry exporter at ZipLogger and every request becomes a waterfall: each operation nested under its parent, across services, with durations and error spans in red. Error logs carry the trace id, so one click takes you from a log line to the exact failing span in the request that produced it.
ZipLogger speaks the Model Context Protocol, so Claude Code, Cursor, and friends can search your logs, cluster error patterns, walk trace waterfalls, and read regression analyses directly. Ask "why is checkout failing in prod?" and your assistant investigates for you.
Fields, tags, releases, commit SHAs, machine and environment, all captured automatically by the SDKs, stored as first-class data, and searchable without any parsing rules.
A native OTLP/HTTP logs receiver accepting protobuf and JSON (gzip included), with resource attributes mapped to first-class fields. Point your existing exporter at ZipLogger and go.
.NET (ILogger, Serilog), Python, Node.js (Pino, Winston), Go (slog), Java, and the browser: all dependency-free, all with the same batching, retry, backpressure, and enrichment behavior.
Everything the UI does, the API does: NDJSON ingestion, search, dashboards, alerts, billing, usage. Per-tenant API keys with one-way hashing and instant revocation.
Multi-tenant from the ground up with Admin, Editor, and Viewer roles, complete separation between workspaces, and audit logs of every destructive action on Team plans and above.
JWT auth with refresh-token rotation, hashed API keys, role-based authorization, tenant isolation, audit trails. Your data is encrypted in transit and at rest, isolated to your workspace, never sold and never used to train models.
Ingestion never blocks or slows your applications, even under burst load, and search stays fast whether you keep thousands of events or billions. Performance you never have to think about.
The free plan includes search, patterns, dashboards, and alerts. Trials of Pro and above unlock AI analysis and git regression detection.