Why a khutbah breaks when you cut it by the clock
Most clippers slice on a timer: sixty seconds, then the next sixty. A reminder does not work that way. The condition arrives before the ruling and the exception after it, so a cut in the wrong place leaves a half-sentence saying something the speaker never said. DeenClipped transcribes the talk with Whisper at word level, then a self-hosted model scores candidate moments and gives its reason — complete ending, question hook, stands alone. Clips are cut on those moments. Nothing is cut to fill an interval.
Recitation is matched to the ayah, not guessed at by ear
When a shaykh recites mid-talk, a phonetic transcript of the Arabic is not a caption — it is a mistake burned into the frame. Recited Quran is matched against a corpus of all 6,236 ayat and drawn as the verse itself with its translation beneath, set in Amiri. The Quran Recitation template goes further and captions scripture and nothing else, so a half-heard aside never sits under a verse. Arabic that is not scripture is captioned in Arabic with an English line under it, from a second translate pass.
Nothing leaves until you have watched it
An Islamic lecture clipper carries a risk a general one does not: a verse misquoted, or a ruling published without its condition. So every clip waits in a review queue. You watch the rendered file — the same bytes that would post — with its score and the model's reasons beside it, then approve, reject or send it back; keyboard shortcuts move you through a stack. Clips containing scripture are flagged QUOTE_RISK and forced into human review. That gate never bypasses, even under automation.
You are charged for six minutes, not forty
A khutbah runs forty minutes and the part worth clipping is often six. Pick a start and an end time and only that stretch is downloaded and processed — one token per source minute, charged on what you selected rather than the file's length. Thirty seconds is the minimum range. Cutting more clips from the same lecture costs nothing further, and neither does re-rendering one or changing its template. Failed renders, and clips you reject before export, are never charged. Basic is free: forty tokens over a seven-day trial.
What happens to a clip after you approve it
Face detection crops the 16:9 source to vertical 9:16 and keeps the speaker in frame. Captions burn in with libass, timed to the words — word-by-word, karaoke, phrase or stacked lines, across five templates. A nasheed is mixed under the speech and ducked, unless you switch it off for the job. Approved clips take posting windows, four a day and eight on Studio, and go to your connected YouTube, TikTok, Instagram and Facebook accounts. Each destination reports its own state, so one platform refusing does not mark the clip failed. There are no view counts here — no connected platform sends them.
Questions
Can it handle a lecture that switches between English and Arabic?
Yes. With no language pinned, transcription runs multilingual and switches per segment, so an English talk containing Arabic recitation is handled instead of having one language forced across the whole file. Spoken Arabic is captioned in Arabic with an English line beneath it, from a second translate pass. You can pin English, Arabic or Urdu instead if you already know what is on the recording.
What happens when the speaker recites Quran?
The recitation is matched against the full corpus of 6,236 ayat and rendered as the ayah with its translation, set in Amiri, rather than as a phonetic guess at the Arabic. The clip is then flagged QUOTE_RISK and forced into human review, so you read the verse on the finished frame before it can go anywhere. The Quran Recitation template captions the scripture and nothing else.
How much of the lecture am I charged for?
Only the stretch you select. Paste a YouTube URL or upload an MP4 or MOV, choose a start and end time, and just those minutes are downloaded, processed and billed, at one token per source minute. Thirty seconds is the minimum. Cutting more clips from the same lecture later costs nothing extra, and clips you reject before export, like failed renders, are never charged. Basic is free: forty tokens over a seven-day trial.
Does it post the clips automatically?
No. Nothing publishes without a human approval — that is the point of the review queue. Once you approve a clip it takes a posting window, four a day and eight on Studio, and goes to your connected YouTube, TikTok, Instagram or Facebook account. Clips containing scripture carry a further gate that automation cannot skip, whatever else is switched on.
Basic includes the whole workflow.
Import a source, generate clips, review every one and publish to your own connected channels. Upgrade when you need more.