WHIP

Inspiration People save food videos all day, but actually cooking from them is still painful. Ingredient lists are hidden in long captions, websites are noisy, and switching between apps breaks flow. We built WHIP to turn recipe links into an immediate, usable cooking experience.

What it does WHIP converts TikTok, Instagram, and web recipe links into structured recipes in seconds:

extracts ingredients, servings, and steps builds a clean shopping list provides a step-by-step cook mode supports demo recipes for first-time users shows cooked-state tracking and clearer completion flow How we built it WHIP is a native iOS app built with SwiftUI + SwiftData. The pipeline:

Detect link type (video/social/web). For videos, fetch metadata/transcription flow. For websites, fetch and clean content before parsing. Send normalized text to our worker-based AI parsing endpoint. Save structured recipe data locally and cache remotely. Render recipe cards, detail view, shopping list, and cook mode. We also added robust fallback handling:

remote cache first, then worker parse local demo fallback if remote demo endpoint fails local image persistence and image recovery flow Challenges we ran into noisy website content (especially recipe pages with lots of UI text) inconsistent social content quality (some links don’t contain real recipe data) onboarding clarity for first-time users under hackathon time pressure balancing polish with reliability near deadline Accomplishments that we're proud of fast end-to-end flow: link -> parsed recipe -> shopping/cooking strong UX improvements under tight time: multi-step onboarding with explanation + demo flow clearer post-import guidance improved cook mode readability completion feedback and cooked badge stable fallback architecture for demos and parsing What we learned onboarding quality directly impacts perceived product quality reliability and fallbacks matter more than adding extra features late small UX fixes (copy, flow, state handling) can dramatically improve retention What's next for WHIP stronger structured time extraction (prep/cook/total from more sources) richer sharing/collaboration flows broader parser tuning for edge-case recipe formats more personalization and meal planning features

Built With

• cloudflare-workers
• gemini-api
• ios-share-extension
• revenuecat
• supabase
• swift
• swiftdata
• swiftui