Surah Al-Fatiha revision check (prototype)
Recite from the start of the surah. Your words appear as you recite. Red = possibly a different word. Amber = unsure or not said. Experimental: it compares the words the model hears with the expected words. It does not judge tajweed or pronunciation, and it can be wrong.
Test prototype - not a finished app. Al-Fatiha only. First load downloads a speech model of roughly 140 MB (then cached), so use Wi-Fi. Word updates can lag several seconds behind your voice, so it is near-live, not instant. It can wrongly mark a correct word, and it can miss real mistakes. Detection of vowel-mark mistakes has NOT been validated. It does not judge tajweed or pronunciation. Do not rely on it for your Quran learning.
Privacy: speech recognition runs in your browser, and the page does not upload or save your audio. Loading it fetches code from cdn.jsdelivr.net and the model from huggingface.co, so those services can see your IP address.
Privacy: speech recognition runs in your browser, and the page does not upload or save your audio. Loading it fetches code from cdn.jsdelivr.net and the model from huggingface.co, so those services can see your IP address.
Loading the speech model (about 75 MB, once). Audio stays on your device.
● right● possible different word● unsure / not said
Incorrect words
Speech model: tarteel-ai/whisper-base-ar-quran (Apache-2.0), run in your browser. Quran text: Tanzil Project (tanzil.net), CC BY 3.0, unaltered. Nothing you recite is recorded or uploaded. This is an independent prototype, not affiliated with Tarteel.