DeveloTools

Budget Calc

Set a performance budget by asset type and see estimated download time per category.

100% client-side

Profile & Budget

Enter a budget (KB) per category, and optionally your page's current size to compare against it. JS costs more CPU per KB than images or fonts to parse and execute — see “How it works” below.
JavaScript
CSS
Images
Fonts
Other

Estimated impact

—
Total download time for the full budget
This is a rough download-time estimate, not a lab measurement. Real FCP/LCP also depends on request order, render-blocking resources and parse/execute time.
Budget Current (within budget) Current (over budget)

How the budget estimate works

Download time is estimated the same way network throttling tools do: convert the byte budget to bits, then divide by the connection's bits-per-second throughput.

time (seconds) = (budget_KB × 8) ÷ throughput_Kbps

Example: 500 KB total budget on a Slow 4G profile (~400 Kbps)
time = (500 × 8) ÷ 400 = 10 seconds

Byte size alone doesn't tell the whole story, though. Per the RAIL performance model, JavaScript is disproportionately expensive versus images or fonts of the same size: images just need to be decoded and painted, while JS has to be parsed, compiled and executed by the main thread before it does anything — on a mid-range mobile CPU, that processing overhead per KB is significantly higher than for other asset types. That's why real-world performance budgets typically cap JS more tightly than image weight, even at equal byte size.

Frequently asked questions

Why does JavaScript cost more than its file size suggests?

Unlike images or fonts, JS bytes have to be parsed, compiled and executed by the device's CPU before anything happens — on a mid-range mobile phone that processing cost per KB is much higher than for images, which is why performance budgets usually cap JS more tightly than image weight even at the same byte size.

Where do the network speed presets come from?

They're rounded, representative throughput figures for common connection classes (Fast 4G, Slow 4G, 3G), similar to the presets used by browser dev tools and Lighthouse-style throttling. Treat them as approximate, not a guarantee of any specific carrier or device.

Is the impact-timing estimate a real lab measurement?

No. It's a rough model based only on download time for the bytes you entered, assuming a single sequential connection. Real page-load metrics like FCP and LCP also depend on request order, render-blocking resources, parse/execute time and other requests — use a real profiler (like Lighthouse) for lab-accurate numbers.

Can I compare my current page's size against a budget?

Yes. Enter your page's current size per category next to the budget, and the chart shows which categories are over or under budget.

Related tools