Qur’an Explorer · The unified platform
Qur’an Explorer brings Sedrah AI’s grounded question answering, recitation feedback, and evidence-linked visualizations into one study environment. Every path leads back to the Arabic text, its source, and the limits of what the system can claim.
Platform features
Search, audio, visualization, and progress tracking are different views over the same source material—not disconnected tools.
Ask in Arabic or English, search by meaning, or enter a verse reference. Retrieved āyāt appear before AI-assisted commentary so the source can be inspected directly.
Compare a learner’s recording with a reference recitation, inspect word timing, and identify where pronunciation needs attention.
Follow a page as each word is recited, replay an individual word, or hide the text for memory practice.
Move from a broad textual pattern to the exact verses behind it, while keeping computation, translation, and commentary visibly separate.
Review recitation attempts, pronunciation scores, studied āyāt, and improvement over time in a private learner history.
Overall architecture
The browser talks to one gateway. Behind it, independently deployable services handle content, retrieval, generation, speech, Tajweed, and learner accounts.
Surahs, āyāt, Tafsīr, audio, and Muṣḥaf layout.
Exact references, word counts, and multilingual semantic search.
AI-assisted commentary constrained by retrieved evidence.
Streaming transcription, alignment, and word-level feedback.
Recitation embeddings compared with a reviewed reference index.
Sign-in, attempts, study history, and progress summaries.
A typical question flows from Explorer → Gateway → Retrieval → Generation, then returns with verse citations. A recitation streams through Gateway → Pronunciation or Tajweed and returns aligned, word-level results. No service reaches into another service’s database.
Visualization examples
Each view answers a different study question and lets the learner move from overview to detail.
Select a Name, follow a represented co-occurrence, and inspect the Arabic verse evidence. The interface labels corpus output as computation—not interpretation.
Open the interactive mapExplore all 286 verses through t-SNE or PCA, filter six experimental thematic classes, reveal the anchor verses, and inspect the Arabic evidence in a searchable table.
Open the semantic mapOverlay word boundaries on a waveform and spectrogram so a score can be traced to timing, alignment confidence, and reference audio.
Turn individual practice attempts into a legible history of consistency and improvement without ranking the learner against others.
Trust & sacred-text safeguards
Sedrah AI supports exploration and study. It does not replace the Muṣḥaf, qualified scholarship, Tafsīr, memorization, or instruction from knowledgeable teachers.
Arabic text remains distinct from normalized search text, with verse boundaries and provenance preserved.
Counts, clusters, similarities, and model scores are reproducible outputs with explicit methodological limits.
Reviewed commentary is labeled by source, reviewer, scope, and revision status.
Generated assistance links back to retrieved evidence and is never silently presented as Tafsīr.