Justlocal.ai

The problem it solves

**Problem: **Finding truly authentic and relevant local recommendations is a persistent challenge. Existing platforms often rely on static, structured data that can be outdated, overly commercialized, or simply miss the vibrant, real-time pulse of a city. Users are often left sifting through generic reviews or sponsored content, struggling to find places that genuinely resonate with local experiences and current trends. Imagine planning a trip or even just a night out, and feeling unsure if the "top-rated" restaurant is truly a local gem or just has a great marketing budget. This disconnect leads to missed opportunities and less fulfilling local experiences.

Our Solution - LocalGuru with Farcaster Integration: LocalGuru bridges this gap by offering a hybrid location discovery experience. We don't just give you the "official story" from structured databases; we enrich it with the authentic, real-time voice of the community through Farcaster.

Challenges we ran into

Challenge 1: Navigating Neynar API Nuances & Limitations Obstacle: The Neynar API, while powerful for accessing Farcaster data, has its specific constraints, such as strict rate limits (e.g., 1-10 results for certain queries) and evolving endpoint behaviors. Early on, we encountered 404 errors, "ExceededMaxLimit" issues, and "Route not found" errors, which hampered our ability to reliably fetch relevant social data. Simply querying with combined search terms often yielded zero results due to the specificity. Our Solution: Multi-Strategy Robust Search & Fallbacks: We didn't just fix the endpoints; we engineered a sophisticated, multi-layered search strategy within our FarcasterClient.

**Challenge 2: **Reliable Location Extraction from Unstructured Social Text Obstacle: Farcaster posts are free-form text. Reliably identifying and extracting specific, geocodable location names from this unstructured data is a significant NLP challenge. Our initial API-based location extraction (/api/extract-location-names) occasionally faced 500 errors or couldn't extract locations from more nuanced cast texts. This meant even if we found relevant casts, they wouldn't appear on the map or in the location list. Our Solution: Graceful Degradation & Multi-Source Location Fallbacks: We implemented a robust fallback system for location data