Everything you need to catch visual regressions before they reach production.
Golden Record
Keep one approved baseline of screenshots per project and compare every new run against it to catch regressions. Approve intentional changes to fold them in, or let a baseline branch update the record automatically. See the comparison docs for the full workflow.
ꟻLIP Comparison
Use NVIDIA's ꟻLIP algorithm to evaluate differences the way a human eye would perceive them. Filter out changes that are technically present but visually imperceptible, and focus on the ones your users will actually notice. See the difference evaluators docs.
Pixel by Pixel Comparison
Detect every single-pixel change between screenshots with an exact XOR comparison. Ideal for catching regressions that subtler methods might miss, from shifted elements to color value drift. See the difference evaluators docs.
Metadata Based Comparison
Tag each test run with metadata like platform, resolution, or graphics API, and let PixelEagle automatically pair runs for comparison. No manual selection needed - just upload and let metadata matching do the work. See auto comparison by metadata in the docs.
Software Forge Integration
Feed comparison results back into your CI pipeline so pull requests show pass/fail status based on visual diffs. Upload screenshots from CI with the PixelEagle CLI and get automated feedback on every commit. Follow the CI integration guide.
Sensitivity Settings
Set a difference threshold per project to control how much change counts as a regression. Tune it to ignore sub-pixel rendering variations or anti-aliasing noise while still catching meaningful visual changes. See thresholds and tolerance in the docs.
Screenshot History
Track how any screenshot evolves over time across runs. Select a screen name and browse every version side by side, with full visual comparison between consecutive changes. Quickly spot when a regression was introduced and how the UI has drifted.
Public Projects
Make any project publicly visible so contributors, QA teams, or the open-source community can browse screenshots and comparison results without needing an account.
Coming Soon
We're building more tools to make visual testing even better.
Masked Zones
Define regions to exclude from comparison, such as timestamps, ads, or dynamic content, so expected variation does not trigger false positives.
Encrypted Projects
Screenshots are encrypted at rest with per-project keys. Stolen backups or leaked storage snapshots stay unreadable, so your visual data is protected even if the underlying storage is compromised.
Placeholder Detection
Automatically flag common development artifacts in your screenshots, such as placeholder textures, missing assets, or checkerboard patterns, before they ship to production.