Islamic Audio & Bayan Transcriber
Upload an audio file and Qaaf AI produces a clean transcript, then translates it if you need it. It is built for the case general transcription tools fail: a scholar speaking Urdu who quotes classical Arabic mid-sentence, where the Quranic and Hadith references have to survive intact.
Try it free — no account needed
Upload a file and see the quality for yourself. You get 100 free credits right now, with no sign-up and no card. When they run out, a free account gives you 200 more.
Processing...
Guides for this workflow
All guides →- The Best AI to Transcribe Urdu Islamic Lectures to English in 2026
- How to Extract Urdu Text from Mufti Tariq Masood's YouTube Bayans
- Generating English Text for Urdu Islamic YouTube Videos
Browse every language pair on the supported languages page.
Frequently asked questions
Can it transcribe Urdu lectures that quote Arabic?
Yes, and that is the case it was built for. A bayan is usually Urdu speech carrying classical Arabic quotation inside it. Mainstream speech models are trained on conversational English and lose the Arabic, or render it phonetically as Urdu, which destroys the reference.
Can I transcribe a lecture that is only on YouTube?
Save the audio first and upload that. Links are not fetched for you any more; any audio file in MP3, WAV, M4A or OGG is accepted, including a recording saved from a video.
Can it translate the lecture as well as transcribe it?
Yes, in the same pass. You get the transcript in the language spoken and a translation into your chosen target language, so an Urdu bayan can come back as English text. Subtitles can be produced from the same transcript for video.
How much audio can I transcribe free?
About 30 minutes, with no account and no card needed. That is enough for a full lecture, so you can judge the accuracy on your own recording rather than on a demo we chose. Registering adds more, and for a whole series of lectures it is worth asking us on WhatsApp first.
What audio quality does it need?
Ordinary lecture recordings work, including phone recordings from the back of a hall. Heavy background noise, overlapping speakers and very low bitrate audio reduce accuracy. Where a passage cannot be made out it is marked rather than guessed at.