Quran Audio Recordings With Verse And Word Annotations

187,080 verse-level audio samples from the Quran, organized into specialized datasets for verses and individual words. The dataset was created by Buraaq and was last updated in January 2026. It includes multiple recitations with granular annotations.

Use Cases

• Train automatic speech recognition models on verse-level audio samples to transcribe Quranic recitations.
• Develop forced alignment models using individual word annotations to map audio segments to specific text tokens.
• Analyze prosodic features across multiple reciters using the annotated audio recordings.
• Build text-to-speech systems for Quranic Arabic by learning from the provided recitation samples.

Strengths

• Contains 187,080 audio samples at the verse level.
• Offers granular annotations down to the individual word for precise audio-text alignment.
• Includes multiple recitations of the complete Quranic text.

Limitations

• Specific audio quality metrics, sample rates, and file formats are not detailed.
• The dataset's size in gigabytes and the exact number of reciters are unknown.
• Potential for limited demographic or recitation style diversity.

Provenance

Source

The Quran (Holy Book of Islam).

Collection Method

Audio recordings of recitations with subsequent annotation.

Freshness

Last updated on 2026-01-27.