🎙️ Quran Multi-Aligner

  • Transcribe and split any recitation by pauses within 1-2 minutes
  • Get precise pause-, verse-, word- and character-level timestamps, exportable as JSON
  • GPU-powered API usage with daily quotas
  • Reliable confidence system to flag uncertain segments and missed words — no silent errors
  • Robust tolerance to noise, speaker variation and suboptimal audio quality, particularly with the large model
  • Not intended for incorrect or fragmented recitations; most suited for correct, continuous recitations (repetitions handled)
  • Feedback and contributions are welcome
ASR Model

Large: more robust to noisy/non-studio recitations but much slower (10x bigger)

Device

Daily GPU usage limits. Unlimitted CPU usage but slower

25 1000
500 2000
0 300