Scale Up Your Snapshot Testing Without the Friction
summary
Snapshot testing at scale means generate images in CI to avoid ARM/Intel and OS drift, store heavy PNGs outside Git (a full-screen snapshot ≈ 50 KB per snapshot, so 1,000 ≈ 50 MB and 100 UI churns ≈ 5 GB), and accept that 99% intentional changes justify an “accept” workflow on PRs. Build artifacts or a hosted service (Screenshotbot is open source and offers PR statuses, Slack alerts, image zoom/diff, history with bisect, masking for animations) to cut flakiness and developer friction.
AI-written summary · about
chapters (10)
- 0:00Introduction to snapshot testing
- 1:49Snapshot vs screenshot vs golden tests
- 4:40Why teams abandon snapshot tests
- 7:06Storage challenges in snapshot testing
- 9:33Solutions to snapshot friction
- 12:19Reducing intentional change friction
- 15:07Screenshot Bot workflow demo
- 16:52Better image comparison tools
- 18:18Advanced Screenshot Bot features
- 20:25Call to action and Q&A
related talks
35mBuilding a screenshot testing pipeline that scales
35mEasy screenshot testing with Compose
31mAn Introduction to Effective Snapshot Testing on Android
58mUnit Testing with Screenshots: Breaking the Sound Barrier
18mScreenshot testing – The practical talk
23mScreenshot testing with ViteShot
31mAutomating UX Compliance Testing
37mSnapshot testing (and more!) with Paparazzi
