Struggling with IP Geolocation API accuracy for city detection?

Hey everyone,

We've been running our web tool, 'What is my City Name', which is designed to provide users with their current city based on their IP address. While the concept is straightforward, we're encountering some significant hurdles specifically concerning city-level accuracy with the various IP Geolocation APIs we've tested. We get pretty solid data for country and even regional identification, but when it comes to pinpointing the exact city, the results are often inconsistent or downright incorrect, particularly when dealing with users behind complex ISP routing or VPNs. This directly impacts the reliability of our core offering and user trust.

I'm really trying to dig deep into strategies for enhancing this. Has anyone here had success combining data from multiple IP Geolocation APIs or integrating external, perhaps less common, datasets to cross-reference and improve the confidence scores? Specifically, I'm looking for methods to intelligently weigh different API responses or even algorithmic approaches to filter out common proxy/VPN noise from legitimate user locations. It feels like there must be a more robust way to achieve high city-level accuracy beyond just relying on a single vendor's database. Any insights or architectural patterns that have worked for you would be incredibly valuable. Help a brother out please...