The accessibility widget your page speed won't notice
Five accessibility widgets measured with Lighthouse, four on one shared test page and accessiBe on live sites. Inclusify added no blocking time.
How much does Inclusify slow down your page?
It doesn't hold it up. On a realistic mobile connection Inclusify added no blocking time and no long tasks, and did less main-thread work than any other widget we tested. It downloads 46.5 KB before a visitor opens it, and nothing more when they do.
0 ms
added blocking time
0
long tasks from the widget
+74 ms
main-thread work, the least tested
46.5 KB
up front, nothing more on open
Mobile, realistic throttling, median of 3 runs.
Next to four other widgets
Four widgets on one page, accessiBe on live sites. Shorter is lighter.
- Inclusify0 ms
- UserWay0 ms
- EqualWeb0 ms
- AudioEye+42 ms
- accessiBelive sites+151 ms
- Inclusify+109 ms
- EqualWeb+134 ms
- UserWay+163 ms
- accessiBelive sites+547 ms
- AudioEye+1098 ms
- EqualWeb25.9 KB
- Inclusify46.5 KB
- UserWay72.3 KB
- accessiBelive site234.6 KB
- AudioEye297.9 KB
- Inclusify0 KB
- UserWay44.1 KB
- AudioEye106.5 KB
- accessiBelive site127.7 KB
Why it stays light
It waits for your page
The main script loads after your page has finished loading, so it never competes with your own images and scripts.
The panel is built on first open
Until a visitor asks for it, only the button exists. Opening the panel downloads nothing more.
Only your language
A site gets its own language in the bundle. Every other language loads only if a visitor switches.
Minified and compressed
46.5 KB on the wire, from one host, in 2 requests.
How we measured
Every figure we measured
All profiles and all metrics, including the ones where another widget comes out ahead.
Mobile, realistic throttling, median of 3 runs
Inclusify
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +74 ms
- Download
- 46.6 KB
- Requests
- 2
- LCP
- +6 ms
- FCP
- -16 ms
- Speed Index
- +16 ms
- CLS
- 0
- Score
- 0
UserWay
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +213 ms
- Download
- 72.3 KB
- Requests
- 9
- LCP
- +1 ms
- FCP
- +2 ms
- Speed Index
- +34 ms
- CLS
- 0
- Score
- 0
AudioEye
- Blocking time
- +99 ms
- Long tasks
- 199 ms
- Main-thread work
- +1406 ms
- Download
- 298.1 KB
- Requests
- 12
- LCP
- -3 ms
- FCP
- +1 ms
- Speed Index
- +55 ms
- CLS
- 0
- Score
- -1
EqualWeb
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +164 ms
- Download
- 25.9 KB
- Requests
- 8
- LCP
- -2 ms
- FCP
- +7 ms
- Speed Index
- +66 ms
- CLS
- 0
- Score
- 0
Mobile, Lighthouse default, median of 7 runs
Inclusify
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +109 ms
- Download
- 46.5 KB
- Requests
- 2
- LCP
- +2 ms
- FCP
- 0 ms
- Speed Index
- 0 ms
- CLS
- 0
- Score
- 0
UserWay
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +163 ms
- Download
- 71.9 KB
- Requests
- 8
- LCP
- -74 ms
- FCP
- -1 ms
- Speed Index
- -1 ms
- CLS
- 0
- Score
- 0
AudioEye
- Blocking time
- +42 ms
- Long tasks
- 184 ms
- Main-thread work
- +1098 ms
- Download
- 298.0 KB
- Requests
- 12
- LCP
- -73 ms
- FCP
- -1 ms
- Speed Index
- +1117 ms
- CLS
- 0
- Score
- 0
EqualWeb
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +134 ms
- Download
- 25.9 KB
- Requests
- 8
- LCP
- -74 ms
- FCP
- -1 ms
- Speed Index
- -1 ms
- CLS
- 0
- Score
- 0
One of the 7 default-mobile runs had an Inclusify long task (122 ms); the median is 0 ms.
Desktop, median of 5 runs
Inclusify
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +69 ms
- Download
- 46.5 KB
- Requests
- 2
- LCP
- +16 ms
- FCP
- +5 ms
- Speed Index
- +5 ms
- CLS
- 0
- Score
- 0
UserWay
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +40 ms
- Download
- 71.9 KB
- Requests
- 8
- LCP
- +10 ms
- FCP
- 0 ms
- Speed Index
- 0 ms
- CLS
- 0
- Score
- 0
AudioEye
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +363 ms
- Download
- 297.9 KB
- Requests
- 12
- LCP
- -8 ms
- FCP
- +1 ms
- Speed Index
- +254 ms
- CLS
- 0
- Score
- 0
EqualWeb
- Blocking time
- 0 ms
- Long tasks
- 0 ms
- Main-thread work
- +68 ms
- Download
- 25.8 KB
- Requests
- 8
- LCP
- -7 ms
- FCP
- +3 ms
- Speed Index
- +3 ms
- CLS
- 0
- Score
- 0
accessiBe on three live sites, with it against without it, median over sites
Added blocking time
- Mobile
- +151 ms
- Desktop
- +6 ms
Widget long tasks
- Mobile
- 217 ms
- Desktop
- 54 ms
Added main-thread work
- Mobile
- +547 ms
- Desktop
- +84 ms
Downloaded up front
- Mobile
- 234.7 KB
- Desktop
- 234.6 KB
Requests up front
- Mobile
- 2
- Desktop
- 2
Largest Contentful Paint change
- Mobile
- +19 ms
- Desktop
- +150 ms
First Contentful Paint change
- Mobile
- +28 ms
- Desktop
- +57 ms
Speed Index change
- Mobile
- +843 ms
- Desktop
- +25 ms
Layout shift change
- Mobile
- 0
- Desktop
- 0
Lighthouse score change
- Mobile
- -3
- Desktop
- -3
Before and after a visitor opens the widget, mobile
Inclusify
- Requests up front
- 2
- Downloaded up front
- 46.5 KB
- Requests on open
- 0
- Downloaded on open
- 0 KB
accessiBe (live site)
- Requests up front
- 2
- Downloaded up front
- 234.6 KB
- Requests on open
- 3
- Downloaded on open
- 127.7 KB
UserWay
- Requests up front
- 9
- Downloaded up front
- 72.3 KB
- Requests on open
- 8
- Downloaded on open
- 44.1 KB
AudioEye
- Requests up front
- 12
- Downloaded up front
- 297.9 KB
- Requests on open
- 2
- Downloaded on open
- 106.5 KB
EqualWeb (panel did not open)
- Requests up front
- 8
- Downloaded up front
- 25.9 KB
- Requests on open
- 4
- Downloaded on open
- 6.8 KB
Page speed changes are the difference from the same page with no widget. Paint timings on live sites move by hundreds of milliseconds from run to run.
Questions about the benchmark.
No. A widget runs in one visitor's browser and changes nothing your site sends, whichever widget it is and however fast it loads. This benchmark measures what a widget costs your page speed, not what it does for accessibility. Barriers are fixed in your markup, and a scan tells you where they are.
Our scope and limitations page says what our own widget and scanner do not do.
Because it would not render on the test page: it downloads and runs its script, then never fetches its site configuration away from a licensed domain. So it was measured on three live sites running it, each loaded as-is and with accessiBe blocked. The method differs from the other four, so compare its figures directionally.
Yes. The test page, every snippet, the scripts that ran Lighthouse and the raw results of every run are kept, and we send them on request: write to support@inclusifyapp.com. The method on this page lists the tool versions and settings, so the run can also be repeated from scratch.
Yes, because every vendor ships new versions of its script, and so do we. These figures are the scripts served on 24 September 2026, and the date is on the page for that reason. We run the benchmark again when our own widget changes in a way that could move its figures, and update this page with the new run.
Raw results and the harness: support@inclusifyapp.com. What a widget can and cannot do: Scope and limitations.
Compared feature by feature: Inclusify vs accessiBe, Inclusify vs UserWay, Inclusify vs AudioEye, Inclusify vs EqualWeb.
Make your website accessible today.
Accessibility opens your site to millions more customers - and protects your business while you do it.
- Scans WCAG
- Scans ADA
- Report in 1 minute