Every bug, explained.

Live inspectors for the data behind the bug, one-tap reports from the device, and AI that answers from your code. All of it runs on your machine.

Inspector

The inspector lives in your browser

Open a tab and your device streams in live: network calls, database rows, crashes, analytics. No IDE, no USB cable.

  • Network: requests and responses with bodies, headers, timing
  • Database: browse and edit Room / SQLite tables live
  • Crashes: stack traces the moment they happen
  • Analytics: every event, in order
  • SharedPreferences and DataStore, editable
chrome-extension://snapbug
Network24 requests
Database3 tables
Crashes1 new
Analytics142 events

Device

WebRTC P2P

Chrome

Scan to pair

Pairing

Pairing by room code, no network setup

Device and browser find each other over a peer-to-peer connection. No IP addresses, no port forwarding, no ticket to IT.

  • A six-digit room code and the device is connected
  • Works across networks: office, home Wi-Fi, VPN

Snapbug CLI

A local daemon next to your repository

Snapbug CLI connects the inspector to your repository and your build tools. AI reads your code locally today; automated fixes and pull requests come later.

  • AI reads your repository locally and explains the cause
  • Build & run: trigger Gradle / Xcode builds from the browser
  • Pull requests: a verified fix becomes a PR (coming soon)
  • Configured from the inspector, runs entirely locally
snapbug analyzetrace · #4218

› > snapbug-cli analyzing crash...

> NullPointerException at UserProfile.kt:42

> Generating fix… (preview)

  1. Apply fix to UserProfile.kt
  2. Run ./gradlew assembleDebug
  3. Create PR #142 with fix
Checkout
↗ overflow
Bug Annotation Preview

Free in every tier

Reports a developer can act on

Capture the bug exactly as it happened. High-fidelity screenshots, recordings, and an annotation editor pin the problem down before any analysis.

  • High-fidelity screenshots and video
  • Arrows, boxes, and text right on the capture
  • Share to Slack or Jira in one tap

AI engine

Analysis from captured data

Snapbug's AI starts from captured data: the screenshot, the failing request, the crash, your code. The analysis starts from facts, not from a ticket description.

Steps to Reproduce

AI writes the step-by-step repro from what the device recorded, instead of a "cannot reproduce" reply.

QA Notes

PRO

Root-cause analysis grounded in your source: what broke, where, and how risky the fix is.

Bug Fix

Coming soon

A ready-to-review diff, built and verified back on the device before anyone opens an IDE. Coming soon.