AI Crash Grouping — Automatic Crash Deduplication for Mobile & Web

Why exact-match grouping breaks

Many crash tools group issues by matching the stack trace as a string. That works until a build gets obfuscated, a method gets inlined, or a line number shifts between releases — at which point the same underlying crash can get filed as a new, separate issue each time the trace looks slightly different. The result is a growing backlog of “unique” issues that are actually five copies of one bug.

How Crash AI groups instead

Crash AI clusters by pattern — the shape of the crash, not a literal string match on the trace — so a rename or a shifted line number doesn’t split one crash into several. Each grouped issue carries a first-seen date, the affected devices and OS versions, and a trend line, so you’re triaging real distinct problems instead of duplicates. In practice this has taken teams from 8,000 scattered reports down to roughly 150 core issues, cutting time spent on resolution by about 95% and duplicate volume by about 80% (published Shakebug figures, see AI bug reporting).

See a grouped crash in the dashboard

Crash AI turns raw crash noise into a short list of grouped issues. Open one and you get the whole picture — occurrences over time, a breakdown by app version, the affected devices and OS versions, and the full stack trace.

Shakebug Crash AI dashboard listing crashes clustered into distinct issues with occurrence counts — Duplicate crashes clustered into a short list of distinct issues.

Shakebug Crash AI crash detail showing occurrences over time, a crash summary by app version, and an app version distribution pie chart for one grouped iOS crash — A grouped issue: first-seen date, occurrences trend, and the affected app-version breakdown.

Shakebug Crash AI crash detail showing the full stack trace under the Description tab for a grouped iOS uncaught exception — The complete stack trace for the same grouped crash, under the Description tab.

How to turn it on

There is no separate crash-reporting module to wire up — crash capture and grouping start as soon as the SDK is initialized on each platform.

Android (Kotlin)

ShakeBug.sharedInstance().initiateWithKey(this, "<Your Key>")

iOS (Swift) — automatic once initialized; optionally attach a short screen recording to crash reports

ShakeBugSDK.sharedInstance.initiateWithKey("<Your Key>")
ShakeBugSDK.sharedInstance.allowAutoCaptureScreenForCrashes(true)

Flutter (Dart) — wrap your root widget; crash capture runs automatically

import 'package:shakebug/shakebug.dart';

ShakebugSDK(
  androidAppKey: 'your_android_appkey',
  iosAppKey: 'your_ios_appkey',
  child: MaterialApp(home: const HomeScreen()),
);

Web (React / Next.js) — crash capture defaults on via the allowCrashReport prop

<Shakebug appKey="YOUR_APP_KEY" bundleId="YOUR_BUNDLE_ID" allowCrashReport={true} />

React Native — crash capture defaults on via the allowCrashReport prop on <ShakebugView>

<ShakebugView allowCrashReport={true}> ... </ShakebugView>

No separate crash-reporting setup step. If you’ve already installed the SDK for bug reporting, crash grouping is already running.

AI Crash Grouping FAQs

What is AI crash grouping?

AI crash grouping automatically clusters duplicate crash reports into a single issue using pattern recognition, rather than matching the stack trace as an exact string.

How is this different from stack-trace matching?

Exact stack-trace matching breaks when a build is obfuscated or a line number shifts between releases, splitting one crash into multiple issues. Pattern-based grouping clusters by the shape of the crash instead, so it survives those changes.

Does Shakebug's Crash AI work across platforms?

Yes. The same Crash AI engine runs across iOS, Android, Flutter, React Native, and web from one SDK.

Do I need to configure anything to get AI crash grouping?

No. Once the Shakebug SDK is initialized for your platform, crash capture and grouping run automatically.

Can I see which devices and OS versions are affected by a grouped crash?

Yes. Each grouped issue shows a first-seen date, the affected devices and OS versions, and a trend over time.

AI Crash Grouping That Survives Obfuscation and Renamed Builds