Our Data

Data Sources

Core restaurant data (name, address, phone, website, hours, ratings, and photos) is sourced from Google Places via Outscraper. This gives us high-quality, frequently updated data consistent with what users already see on Google Maps.

We do not fabricate, invent, or misrepresent restaurant information.Ratings and review counts are sourced directly from Google and linked back to Google Maps for verification.

Regional Cuisine Classification

Cuisine tags are generated by Claude AI (Anthropic) using the restaurant name, description, and known keyword signals. Classification is conservative — we only assign a tag when there is clear evidence. An empty cuisine tag is better than a wrong one.

Tags are pipe-separated and can include multiple values (e.g., a restaurant can be both North Indian and Indo-Chinese). We classify against a fixed taxonomy of 12 regional styles.

Dietary Classification

Dietary tags (Jain, pure vegetarian, vegan, halal, gluten-free) are assigned conservatively based on explicit signals in the restaurant name or description. We do not infer dietary suitability from cuisine type alone.

Always confirm directly with the restaurant before relying on dietary tags for health or religious compliance. Restaurant menus and practices change.

Phased Publishing

We publish restaurants in priority order based on data quality, rating, review count, and cuisine/dietary enrichment completeness. Not all restaurants in our dataset are currently visible. We add approximately 100 new listings per week.

Hidden Gem Algorithm

A restaurant is flagged as a "hidden gem" if it meets all three criteria: rating ≥ 4.5 stars, fewer than 200 Google reviews, and currently operational. This combination identifies genuinely exceptional restaurants that haven't been discovered by mainstream audiences yet.