Identify a recitation
Hear an ayah — from a speaker, a mosque, a video, or your own voice — and know exactly where it is in the mushaf. Recite the phrase you caught and the matcher searches all 6,236 ayat on your device; or let it listen to a recording that is playing and it matches an acoustic fingerprint of the clip, sending only the fingerprint, not the audio. Cited answers are educational support, not a fatwa.
Two ways to identify
Reciting to search: your browser’s speech recognition hears the phrase (in Chrome and Edge the audio goes to Google or Microsoft) until the on-device model has downloaded in the background — after five minutes with the site open and in view, on a device and browser that can run it, never starting while data saver is on or on a connection as slow as 2G — or you choose it; from then on that model transcribes the phrase, and the matching against the full mushaf index always runs on your device. Identifying a recording fingerprints ten seconds of what the microphone hears and sends only that fingerprint; our server matches the landmarks against known published recitations, the same technique music identification uses.
It names the reciter of a matched recording
Every fingerprinted track carries the reciter it belongs to, so a recording match also says whose recording it is. Results are graded confident or likely against measured thresholds, and a clip that matches nothing in the corpus is refused rather than guessed.
FAQ
Which surah is this? Can it tell from audio?
Yes. Point the microphone at the recitation, or recite the phrase you remember, and it names the surah and ayah, then opens the verse with translation and tafsir.
Does it recognise which reciter is playing?
When the clip matches a recording in the fingerprint corpus, the answer names that recording’s reciter. Voices and recordings outside the corpus are refused, not approximated — it does not identify a person’s own reciting voice.
Is the audio uploaded?
For recite-to-search, it depends on the engine. Your browser’s speech recognition hears the phrase until the on-device model is in use, and in Chrome and Edge that sends the audio to Google or Microsoft; on a device and browser that can run it, and never starting while data saver is on or on a connection as slow as 2G, the model downloads in the background after five minutes with the site open and in view (or when you choose it), unless you have said no to it, and from then on the phrase is transcribed on your device, with one tap back. Either way the matching against the mushaf runs on your device. Recording identification sends only an anonymous acoustic fingerprint — a list of numeric landmarks — which our server matches; the recording itself is not sent.
Cited answers are educational support, not a fatwa. Live teaching is not bookable.