ShipLens
Give your website review something concrete to work with.
ShipLens is a local website delivery checker. It opens pages in Chromium, records repeatable evidence and helps you inspect what a visitor will actually encounter before you publish a change.
Inspect the page that actually runs
A scan can find runtime errors, failed resources, visibly broken images, suspicious blank pages and horizontal overflow. Desktop and mobile observations remain distinct. Reports include screenshots and page details so a finding can be traced back to something you can inspect, rather than a vague quality score.
Keep the evidence close
Each run produces its own local report directory with HTML, JSON, Markdown and images. Explicit routes and configured interactions support repeatable checks. A baseline comparison can distinguish new findings from older ones when the scope and coverage match. Incomplete checks are reported separately from confirmed problems.
Work with your existing assistant
MCP tools and the review API let a compatible assistant collect scoped evidence, record cited assessments and revisit saved cases. The assistant provides the reasoning; ShipLens provides browser evidence and the review record. Fresh runs require fresh assessments: an earlier pass is not silently copied onto a new version.
Use checks for what they can prove
ShipLens is not an automatic visual judge or a guarantee that a website is ready for every user. Human and device testing can still be needed. The CLI does not require a model account or API key and does not upload reports. If you connect an AI client, that client can receive the evidence it requests.