NEW

Page Speed

The dashboard has a new Page Speed section: it measures page load speed in a real browser, with a load video, a waterfall and recommendations on what to fix.

What's inside

All of the organization's runs are in one list. Each shows the page, device, deploy and main metrics. Scheduled runs come with bars for the latest passes, so you can see how speed changed from deploy to deploy.

The Page Speed section in the Layero dashboard: a list of runs with page address, device, deploy, LCP, CLS, TBT and the score
The run list. At the top: how many runs are left this month.

A real browser

The page opens twice in Chrome, just like for a visitor. The run records a video, every request and the main thread's work.

Phone or desktop

Network and CPU speed like a real visitor's, the same for the video and for Lighthouse.

Comparison with the previous run

A new run is compared with the previous scheduled one. The "got worse" mark appears only when the difference exceeds measurement noise.

Schedules and Telegram

On Pro a run starts by itself every 3 or 6 hours, daily or weekly, and the result arrives in Telegram.

A run in two clicks

Pick a project, a branch and a page — with query parameters or on a connected domain if you like. Then a device. The report is ready in about two minutes, and the progress is shown right in the list.

The Measure speed dialog: project, branch, page address and device — phone or desktop
Phone: 37/8 Mbit/s internet, 50 ms latency, CPU 2× slower. Desktop: 90/90 Mbit/s, 20 ms, no slowdown.

Report

At the top: the main load metrics — LCP, TBT, CLS, FCP and Speed Index, each on a scale with Google's thresholds. Below: what to fix first and how much time it saves.

A run report: device, branch, deploy and the main load metrics with colored scales
The report can be copied for an AI assistant, and the request log can be downloaded as HAR.

Load video

This is how a visitor saw the page: frame by frame, with marks for first paint, the main block and visual readiness. You can see how many seconds the screen stayed blank.

Load video tab: a timeline with FCP and LCP marks and a frame of the page on a phone
The timeline can be scrubbed, and the frame follows it.

Waterfall

Every network request on a timeline, with event marks and long tasks of the main thread. This is where you see which file holds up everything else.

Waterfall tab: page requests by file type on a timeline with LCP and Load marks
Requests can be filtered by URL. Next to it, the Resources tab shows page weight by file type and the heaviest files.

Lighthouse and Coach recommendations

Two sets of checks. Lighthouse covers performance, accessibility, best practices and SEO, and every suggestion shows the expected savings. Coach adds privacy checks and explains in plain words what's wrong and in which files.

Lighthouse performance recommendations: document request latency, layout shift causes, render-blocking requests
Lighthouse: what to fix and how much time it saves.
Coach recommendations: async image decoding, long main-thread tasks, redirects, JavaScript size
Coach: every suggestion lists the files it applies to.

Got worse? You'll see it right away

A new run is compared with the previous scheduled run for the same page and device. You can also select several runs in the list and open them side by side.

Two runs of one page compared: metrics side by side, changes marked with color
When comparing, the earliest run is the baseline.

Scheduled runs

On the Pro plan a run can start by itself: every 3 or 6 hours, daily or weekly. Runs of one schedule are grouped in the list under its name. With notifications on, the result of every run arrives in Telegram: what changed, which deploys happened and what to fix first.

The New schedule dialog: page, device, frequency, group name and the Telegram notifications switch
At the bottom of the dialog: when the next run happens and how many runs a month the schedule uses.

From the terminal and from an AI agent

Start a run with the CLI: npx layero@latest perf check --device mobile --path /pricing --wait. The result is the same as in the dashboard. AI agents start a run with the MCP tool check_performance.

Measure your site

Page Speed is available on both plans. Free gives 3 runs a month, Pro gives 500 runs a month and scheduled runs.