Step 1: Configure Your Analysis
Analysis Settings (Research Mode)
Strict mode prevents false positives (e.g. "ama" in "samana"). Compound mode segments sandhi forms like "vatapittau".
Add a New Concept
Save/Load Map
Generated JSON
Step 2: Input Shloka Text
Dashboard
Results
| ID | Shloka | Top Concept | All Scores | Matched Keywords | Chars | Words |
|---|
Research-Grade Matching (v2.0)
ShlokaAI v2.0 uses three matching modes designed for Ayurveda text mining:
- Strict (default): Unicode-aware word boundaries + exact token matching. Prevents false positives like matching "ama" inside "samana".
- Compound: Dictionary-based segmentation for sandhi/samasa forms (e.g. finds "vata" and "pitta" in "vatapittau").
- Legacy: Original substring matching — useful for comparison only, not recommended for publication.
For full morphological decompounding (all sandhi rules), integrate external analyzers such as the INRIA Sanskrit Reader or Sanskrit Heritage API in future pipeline stages.
How to Use ShlokaAI
- Load templates or define concepts with Devanagari/IAST synonyms.
- Choose matching mode and false-positive guards under Analysis Settings.
- Paste or upload text, run analysis, then export wide CSV (per-concept columns) or JSON with full metadata.
Cite this Tool
Varshney, S. (2025). ShlokaAI: The Smart Sanskrit Analysis Platform (v2.0.0). Zenodo. https://doi.org/10.5281/zenodo.17506829
About the Creator
Created by Sparsh Varshney, BAMS Student and Ayurvedic Researcher.
ORCID: 0009-0004-7835-0673
Website: Amidha Ayurveda
GitHub: sciencewithsaucee-sudo/ShlokaAI
Contact: sparsh@amidhaayurveda.com