Live Khutbah Translation: How It Works in Real Time

You are sitting in the masjid on a Friday. The imam is six rows in front of you, speaking fluent Arabic for twenty minutes, and you follow roughly one word in five. You know the khutbah is about something. You know there was an ayah in there, and probably a hadith, and a story. You walk out having prayed but not having understood. This is the problem live khutbah translation is built to solve: put the phone on your lap, watch the Arabic transcribe in real time, and read the translation in your language as the khatib speaks.

This post explains what happens between the imam's voice and the English (or Urdu, French, Turkish, Indonesian) sentences scrolling on your screen: why real-time Arabic translation is harder than real-time English, what the system gets right, what it gets wrong, and when to trust it.

Who actually needs live khutbah translation

Two groups of people, mostly. The first is reverts and born Muslims who never learned Arabic beyond the prayer. In the United States, the United Kingdom, Canada, Australia, much of West Africa and South Asia, and increasingly western Europe, this describes much of the congregation. The khutbah happens in Arabic either fully or for the opening and the ayah recitation, and the translation afterwards (if there is one) is a summary, not the khutbah itself.

The second is travellers. You are in Istanbul, Jakarta, Casablanca, or Dubai for the week, you walk into whichever masjid is closest at 1:15pm on Friday, and the khutbah is in a language you do not share with the local congregation. Arabic is the common denominator. Real-time Arabic translation closes the gap.

A third, smaller group: students of Arabic who can follow most of the khutbah but miss a few terms per minute and want a running safety net. Live translation is useful here as a dictionary that keeps pace with the speaker.

Why real-time Arabic translation is harder than English

Most real-time translation demos you have seen are English to something, or something to English. English ASR has been under relentless attention for fifteen years, trained on tens of thousands of hours of clean speech. Arabic does not have that luxury, and khutbah audio is worse than the Arabic the research datasets are built on.

Three layers of difficulty stack up inside a single sermon.

Three kinds of Arabic in one khutbah

The khatib is not speaking one register. A typical twenty-minute khutbah contains:

• Classical Arabic. Quranic verses quoted verbatim, plus hadith in their original wording. Vowels fully marked, archaic vocabulary, syntactic patterns an MSA speaker does not use in conversation.
• Modern Standard Arabic. The body of the khutbah, the actual sermon content, delivered in the formal register used on Arabic television and in news broadcasts.
• Dialectal asides. When the khatib tells a story or makes a joke, he drops into his own dialect: Egyptian, Levantine, Gulf, Moroccan Darija, Sudanese. Sometimes he stays in dialect for a minute before climbing back into MSA.

Speech recognition trained on Modern Standard Arabic handles the middle layer well. On the Classical Arabic of Quranic quotation, the same systems struggle more, because the phonetic patterns and vocabulary are far older than most of their training data. On dialects such as Moroccan Darija, error rates are higher still, even on clean isolated speech, let alone a speaker swapping in and out of dialect mid-sentence.

Code-switching

Many khatibs, especially in diaspora masjids, switch languages. A khutbah in the United States might open in Arabic, translate the ayah into English, give a story in English, then close in Arabic. A UK khutbah can splice Urdu. A Paris khutbah can splice French. RecitID's Live mode is built for Arabic: it transcribes speech as Arabic, so long passages in another language are not transcribed reliably.

Masjid acoustics

Khutbahs are delivered in large rooms with hard surfaces and carpeted floors. Reverb smears consonants. Low-frequency hum from fans, air-conditioning and electrical systems masks bass frequencies. Children cry, phones buzz, latecomers whisper. The microphone on your phone is a few metres from a speaker that is itself amplified through a PA, so you are recording a reverberant re-radiation of the original voice, not the voice itself. Each of these raises the word error rate.

The two-stage pipeline

Live khutbah translation, like most production speech translation, runs as a cascade: the audio is first transcribed into Arabic text, and that text is then translated. Rather than waiting for the end of the sermon, the app works in short slices of a few seconds, so the Arabic and the translation keep pace with the khatib.

Stage 1: Arabic transcription

Your phone records the khutbah in short chunks of a few seconds, ending each one at a natural pause where it can, so words are not cut in half. Stretches of silence are recognised and not sent for transcription.

Each chunk is cleaned up on our servers (low rumble filtered out, volume levelled) and transcribed in the cloud with Arabic set as the language, using the end of the previous chunk as context so sentences carry across the cut. Each chunk appears on screen as a finished line of Arabic; it is not rewritten afterwards.

Before anything is shown, each transcript goes through checks that catch what speech recognisers tend to produce on noise or silence, such as stock captions or repeated phrases, and drops them.

Stage 2: translation

Each Arabic line is then translated by a large language model, given the preceding lines as context and instructed to keep names, Islamic terms and quotations intact rather than paraphrase them. When the khatib attributes a hadith, the translation carries only the attribution he actually gives.

The translation appears below the Arabic, typically a few seconds after the words are spoken: a chunk has to finish recording before it can be transcribed, and the translation follows right behind the Arabic.

Why not end-to-end speech translation

An alternative architecture skips the intermediate Arabic text and maps Arabic audio directly to English (or your target language) text. This is called end-to-end speech translation. It has a latency advantage (one model instead of two) and can sometimes preserve information that ASR discards, like emphasis and emotion.

The tradeoff is that end-to-end systems are harder to train, harder to debug, and less accurate on code-switched input. For khutbah translation specifically, the cascaded approach has one big win: because the Arabic text is produced explicitly, we can intercept it between stages and work on it. The most important example is how Quran quotations are handled.

Quran-quotation handling

The khatib quotes an ayah. The speech recogniser transcribes it as Arabic text. A general-purpose translator would then treat it like any other sentence, and the result may resemble the ayah but is neither exact nor dignified.

The fix is to stop treating the Quran as ordinary speech. The text of the Quran is fixed and finite, so a system that recognises when it is hearing a quotation can reach for the established translation instead of improvising one. That is what RecitID does during a khutbah: when the khatib recites, the verse is identified and shown with its surah name and an established translation in the reader's chosen language, marked as verified: the words of a recognised translator, not a machine paraphrase. Where no published translation is available in your language, the verse's exact Arabic text is machine-translated instead.

Hadith are recognised and labelled when the khatib introduces them, but they are translated live like the rest of the sermon, because the hadith corpus is far larger and its wording varies between narrations. For the Quran, quotation detection works well once enough of the verse has been recited and transcribed; a short fragment on its own is often not enough to identify the ayah with confidence.

The result: when the khatib quotes enough of the passage containing inna ma'a al-usri yusra, you see the established translation (Saheeh International in English: "Indeed, with hardship [will be] ease", Quran 94:6) rather than an improvised one.

How accurate is it, really

Accuracy varies by three factors that are independent of the app: the acoustics of the room, the dialect and diction of the khatib, and the ambient noise. Here is what to expect.

• Clean MSA, well-amplified, quiet masjid. Transcription is mostly correct. Translation reads well. Quranic quotations are caught and shown with an established translation. This is the case the system is built for.
• Heavy reverb. Large masjid with a high dome or marble interior. Transcription degrades, translation degrades proportionally. Sitting closer to the imam or to a PA speaker helps more than anything else.
• Thick dialect. Moroccan, Algerian, deep Egyptian. Dialectal segments will be transcribed with higher error rate. The MSA portions of the same khutbah remain fine. Quran quotations still get caught because the ayah text is the same regardless of dialect.
• Thick non-Arab accent. A khatib whose first language is Urdu, Turkish, or Malay and who pronounces Arabic with a strong substrate accent. Usually works better than dialect because the underlying phonology is closer to MSA, just coloured.
• Noisy masjid. Crying children, HVAC, latecomers shuffling. More lines come out garbled or are dropped, but the transcript usually recovers once the noise passes.

Put plainly: if you are in the back row of a 2,000-person masjid with a thick reverb and the khatib is speaking colloquial Moroccan, the translation will be noticeably worse than if you are three rows back in a 200-person masjid with a clear PA and a Cairo-trained khatib. The feature is a tool, not a miracle.

Supported languages

Live Khutbah covers 52 translation languages: English, French, Spanish, German, Turkish, Urdu, Indonesian, Malay, Bengali, Russian, Chinese, Japanese, Korean, Portuguese, Swahili, Hausa, Somali, Hindi, Persian, and more.

One difference trips people up. Live Khutbah and the reading translations are two separate systems that happen to cover a similar number of languages. Live Khutbah translates spoken Arabic line by line, as each part of the sermon is transcribed. The reading translations that appear next to detected ayat are human-produced authoritative texts (Saheeh International for English, Hamidullah for French, and so on). Both give you the meaning in your language, but one is a real-time transcription of live speech and the other is a vetted canonical translation of the Quran, so the two language lists are not identical.

Privacy: what happens to the audio

The short version: audio is sent to our servers in short chunks during the session and deleted after it is transcribed; we do not keep the recording. Our AI providers may retain request content for a limited period (see the privacy policy). Your transcript and translation are saved as a session on your device, where you can re-read or delete it.

Real-time ASR at this quality currently requires cloud inference. The models are too large to run on phone hardware without melting the battery in ten minutes. We watch the on-device model space closely, and if a model ships that can do streaming Arabic ASR on an iPhone or Pixel at acceptable quality, we will move there. Until then, audio is processed in the cloud and deleted once it is transcribed. Full retention windows are in the privacy policy.

How to use it in the app

Three steps.

1. Open RecitID and switch to Live Khutbah mode. Select your translation language from the list.
2. Tap record when the khatib starts speaking. Place the phone face-up on your lap or on the carpet in front of you. The built-in phone mic is fine.
3. Watch the Arabic and the translation appear in parallel. The session is saved on your phone as it runs, so you can re-read the khutbah later.

Works for any Arabic speech, not only Friday khutbahs. Islamic lectures, halaqas, taraweeh du'a, weekend seminars, online classes. Anywhere the speaker is speaking Arabic and you want the meaning in your language.

What pairs well with Live Khutbah

Two other features cover cases Live Khutbah does not.

AI Explain takes a specific ayah and gives you a plain-language explanation grounded in Tafsir al-Jalalayn, with its main themes and the source named. Useful after the khutbah when you want to go deeper on a verse the khatib quoted. Pair it with Asbab al-Nuzul if you want to read about the classical circumstances-of-revelation literature.

AI Chat answers free-form questions about what you just heard. "The khatib mentioned three conditions for a valid repentance, can you list them with sources?" That sort of thing.

For the separate problem of identifying the Quranic recitation itself, RecitID Detect matches recited audio against the Quran text and returns the surah, verse, and reciter. Detect and Live Khutbah are different tools: Detect is for recitation (closed text, fixed corpus), Live Khutbah is for sermons (open speech, MT). The explainer on why Shazam can't identify Quran recitation goes into the difference in more detail.

Pricing and limits

Live Khutbah is mainly a paid feature (your first khutbah is free) because cloud transcription and translation cost real money per minute of audio. The pricing page has the current numbers; at time of writing, Monthly Pro includes 5 hours of Live Translation per month, and Annual Pro includes 60 hours per year (boosted to 70 during Ramadan to cover the extra taraweeh and lecture sessions). Extra hour packs are available if you burn through the quota.

Sixty hours is enough for every Friday khutbah for a year (about 22 hours at 25 minutes each), plus a weekly lecture, plus travel. The Ramadan bump exists because heavy users, reverts especially, tend to attend a lot more talks in that month.