Who this is for
Teams investigating slow main-content loading, unstable layouts or heavy assets on important mobile pages.
Page Speed Automation turns a slow page into a testable change.
Page Speed Automation runs a mobile Lighthouse lab check on a page and saves the performance, accessibility, best-practice and SEO scores with the top opportunities.
In buildStatus today · Agent: Diane
Part of the Website Optimiser module. Part of this works today and part does not. The scope below says which is which.
Practical purpose
Mobile Lighthouse lab checks with the top opportunities.
Turn a loading or layout problem into a testable change. Lab results are not real-user measurements, and fixes need verification.
For Next.js codebases, an experiment loop is built that tries one code change at a time on a separate branch, rejects regressions and never deploys. It has not yet run for a client. Real-user Core Web Vitals need field data and are in build. A single lab run cannot establish how every visitor experiences the site.
Service scope
Is this the right fit?
What to bring
Representative URLs, device and connection assumptions, deployment access and field data where available. Record the test conditions before comparing results.
What you receive
Mobile Lighthouse lab checks with the top opportunities and, for Next.js codebases, an experiment loop that tests changes on a separate branch. Changes need review and repeat testing, and a lab score is not a real-user result.
Illustrative example
A large mobile image may be improved with appropriate dimensions and responsive sources. Verify the selected image and visual quality on the actual page.
How to evaluate the work
Compare the same device, connection and page state. Report lab LCP and layout shift separately from real-user Core Web Vitals.
How it works
The workflow, step by step.
The page is tested on the mobile profile
Headless Chrome runs Lighthouse and saves the category scores, FCP, LCP, Speed Index, TBT, CLS and the top opportunities.
Compare lab results and available field data
Repeat browser tests under the same conditions. Review real-user Core Web Vitals separately when enough field data exists, and never relabel a lab measurement as a field result.
Layout shift is treated as a conversion problem
Not a technical score. A control that moves under a thumb costs you the tap, and taps are leads.
Fixes are tested one at a time
For Next.js codebases, the experiment loop makes one code change per iteration on a separate branch, rebuilds, re-runs Lighthouse and keeps the change only if nothing regresses. It never deploys.
Evidence and scope
Search context and workflow facts.
2,400US searches a month for core web vitals
4 scoresfrom one mobile Lighthouse run on Ruh AI's homepage, 10 September 2026: performance, accessibility, best practices and SEO
Search volume: Google Keyword Planner, United States, pulled 31 Aug 2026. It is historical search demand, not a traffic forecast. Other figures describe the documented Ruh AI record, workflow scope or stated limitations. They are not promised results for your business. Read the Ruh AI evidence and measurement limits.
The system behind it
Website Optimiser Lighthouse check
On 10 September 2026 the Ciny team ran a mobile Lighthouse lab check on Ruh AI's homepage and saved all four category scores. The Next.js experiment loop passed its tests but has not yet run for a client. Real-user Core Web Vitals from PostHog are in build.
- In build: lab check runs on request
- built for Ruh AI first
- approval controls depend on the workflow
Signals in
- Your data
- Your brand pack
- What worked last month
Diane · Website Optimiser Lighthouse check
Agree the workflow
- data and access
- approval and delivery
- checks and limitations
Scope notes
Good to know.
- Lab checks run on request. Code changes are tested on a separate branch through the Next.js experiment loop, and a person reviews them before release.
- It names costly third-party scripts so your team can decide what to keep.
- Real-user Core Web Vitals are in build, and today's results are lab measurements.
Why mobile first?
Because a desktop score can hide problems a phone visitor meets. The Lighthouse check uses the mobile profile, so the result reflects a phone-sized test.
Is a Lighthouse score the same as Core Web Vitals?
No. Lighthouse is a lab test under set conditions. Core Web Vitals describe real visits, and that analysis is in build. We report the two separately.
Will it automatically make my site faster?
It finds and lists the top opportunities. For Next.js codebases the experiment loop can test code changes on a separate branch, and a person decides what ships, because a performance change that breaks a layout costs more than the milliseconds it saved.
Work in practice
Read the evidence. Then talk about your business.
Two stories with different evidence levels. Read each source and its limits.
All case studiesFirst-party measurement
Ruh AI
Organic leads, with the evidence attached.
Weekly organic leads rose from 6 in early July to 20 in the week of 24 August 2026. August paid spend was $0.
Discuss your goals- Weekly organic leads
- 6 to 20
- Evidence
- First-party
Published Ruh AI customer story
Rapid Innovation
A more connected marketing workflow.
Ruh AI reports 37% more leads for Rapid Innovation after introducing AI-supported research, publishing and lead qualification.
Discuss your goals- Reported leads
- +37%
- Source
- Ruh AI story
These studies explain the wider work behind Ciny. They do not establish results for every capability. Live, Available, In build and Planned labels still apply.
Explore Page Speed Automation. Ask what works today.
Part of this works today. We show you exactly which part before anything is agreed.