19 accessibility tools, 8 one-click modes, real sign-language translation, AI image description, and full white-label branding — all live today unless marked otherwise.
No changes to your site's code — every visitor customizes their own experience, and it's remembered for their next visit.
Reads page content aloud as visitors tab or click, at 3 reading speeds.
Lists every heading and link on the page for fast keyboard navigation.
An on-screen keyboard for typing into any field without a physical one.
Scales page text up for easier reading.
Adjusts text boldness for legibility.
Widens spacing between letters.
Adds breathing room between lines of text.
Changes how paragraph text is aligned.
Hover any text to see it enlarged in a floating lens.
An oversized pointer, in 6 color choices.
Adjusts color intensity across the page.
Light and dark high-contrast modes.
Plus independent text, title, and background color overrides, with 16 swatches each — for visitors who need a specific color combination to read comfortably.
Swaps the page's font for one designed to be easier to read with dyslexia.
Dims everything except a band around the cursor, to help track a line of text.
Outlines every link so they're easy to spot.
Pause or fully hide videos on the page.
Mutes host-page video/audio (never the toolbar's own spoken feedback).
Hides non-essential images for distraction-free reading.
Adds a "translate to sign language" button next to page text — see the full sign-language breakdown.
Each bundles a preset of the tools above for a specific need.
Blind and low-vision visitors turn this on (keyboard shortcut Ctrl + /, or a toolbar click) and pick a detail level. Isharat scans every image on the page, has an AI look at each one, and applies a real spoken and written description as the image's actual alt text — so it works with the visitor's own screen reader, not just Isharat's.
alt text is respected by default, with an optional switch to override itDetail-level preferences measured on Beta traffic through 17 August 2026.
If someone already wrote a proper alt description for an image, Isharat leaves it exactly as it is — human accessibility work shouldn't be silently overwritten by a machine. Visitors who want more can flip an optional "Override human alt text" switch, which asks the AI to look at the image afresh, useful when they need deeper visual context or want the text inside the image transcribed.
Beta today — built and working end-to-end, currently running on staging infrastructure rather than promoted across every customer site.
Widget screenshot over an organizational-chart image, showing the three detail-level buttons (Short / Medium / Detailed), the "Override human alt text" toggle in its active state, and a callout on the transcribed Arabic and English text inside the image.
1200 × 750 px · 16:10 · WebP + PNG · EN and AR variantsSign language isn't word-for-word replacement. Isharat analyses the whole sentence to pick the right sign for ambiguous words, then reorders it into real Arabic Sign Language grammar before anything plays.
The engine drops articles, conjunctions, and prepositions that sign language doesn't use; reorders the sentence into natural topic-then-comment structure; and inserts grammatical markers such as انتهى for completed past actions and لا for negation.
Built to accept any Arabic dialect — Egyptian, Gulf, Levantine, Maghrebi — normalise its regional wording, and translate the meaning into standard Arabic Sign Language. It's also built to take text in other source languages and render the meaning in ArSL.
Evaluated on our published 200-sentence, 5-dialect benchmark. Coverage beyond Arabic is under active testing.
Arabic words spelled identically can carry different meanings, and a different meaning can call for a different sign. The engine reads the surrounding sentence to choose, instead of guessing or asking the visitor every time.
Below each sign video there's a three-level comprehension bar. It takes one tap, and it's the most direct quality signal we have — because the people answering it are the people the translation is actually for.
Shown here as a static example — the live version appears under the video inside the widget.
Real Deaf comprehension ratings drive our translation-quality work — which sense the engine picks for an ambiguous word, how natural the ArSL ordering feels, and which gaps get prioritised.
A translation marked 😕 Unclear automatically deactivates its own cached entry. The next person to request that sentence gets a freshly generated translation instead of the one that was just reported as wrong.
Real screenshot of the Isharat sign player with a video paused mid-sign and the three-level feedback bar visible beneath it. Needed in both the English and Arabic widget interfaces.
1200 × 900 px · 4:3 · WebP + PNG · EN and AR variantsEvery part of how the widget looks and presents itself can be customized per subscribing website — so a government ministry or large enterprise gets a widget that matches their own visual identity, not a generic third-party add-on.
Example: a Saudi DGA-styled trigger icon, matching a government portal's own green brand palette.
The same widget, configured three ways. Colors, logos, and trigger styling are set per subscribing site.
Real human interpreters, Egypt's national dictionary, and honest coverage numbers — read the full breakdown.
Explore Sign-Language TranslationTalk to us about a demo, pricing, or a white-label partnership.
Request a Demo