<h2>TL;DR</h2>
OpenPOI is a high-performance, cost-effective alternative to Google Places API. It uses OpenStreetMap (OSM) data powered by a smart 3-layer caching system (Redis + MongoDB) and an automated self-healing worker to ensure millisecond response times.
Try it on RapidAPI: <a href="https://rapidapi.com/blackbunny/api/openpoi-api-global-places-and-poi-data-service" target="_blank" rel="noopener nofollow">OpenPOI API Gateway</a>
<hr>
Overview:
If you’ve ever built an app requiring location data, you know the struggle:
<ul>
<li>Google Places API is incredibly expensive</li>
<li>Raw OpenStreetMap (OSM) data is powerful but painful to parse and query at scale</li>
</ul>
OpenPOI bridges this gap. It’s a production-ready, cached, and developer-friendly API that brings the power of OSM to your apps without the complexity or high costs.
<hr>
<h2>The Engineering Behind It</h2>
To achieve millisecond response times and high availability, OpenPOI uses a Triple-Layer Data Strategy:
<h3>L1 — Hot Path</h3>
<ul>
<li>Redis cache for frequent and repeat queries</li>
</ul>
<h3>L2 — Persistence Layer</h3>
<ul>
<li>MongoDB with 2dsphere geospatial indexing for fast spatial queries</li>
</ul>
<h3>L3 — Source / Failover</h3>
<ul>
<li>Overpass API integration for real-time OSM data fetching</li>
<li>Used only when data is missing locally</li>
</ul>
<hr>
<h2>The Secret Sauce: Self-Healing Background Worker</h2>
This is where OpenPOI really shines.
When a user searches a new or uncached area:
<ol>
<li>
The API responds instantly with available data
</li>
<li>
A Redis Pub/Sub event is fired
</li>
<li>
A background worker:
<ul>
<li>Scans 6 surrounding hexagonal points</li>
<li>Fetches missing POIs</li>
<li>Backfills the database intelligently</li>
</ul>
</li>
</ol>
Result: The next nearby search hits only L1/L2
No repeated external API calls
Faster responses over time
<hr>
<h2>Key Features</h2>
<ul>
<li>
Hybrid Search
Search by coordinates or city / district names
</li>
<li>
Smart Parsing
Converts complex OSM <code>opening_hours</code> into a simple <code>open_now</code> boolean
</li>
<li>
Developer‑First Design
<ul>
<li>Clean JSON responses</li>
<li>Rich metadata (phone, website, hours)</li>
<li>Categorized and normalized tags</li>
</ul>
</li>
</ul>
<hr>
<h2>🛠️ Tech Stack</h2>
<ul>
<li>Backend: Node.js, Express</li>
<li>Database: MongoDB (GeoSpatial), Redis</li>
<li>Process Management: PM2 (Microservice architecture)</li>
<li>Data Source: OpenStreetMap / Overpass API</li>
</ul>
<hr>
<h2>Get Started</h2>
OpenPOI is live on RapidAPI with a generous free tier for the community.
Full Documentation &amp; API Access: <a href="https://rapidapi.com/blackbunny/api/openpoi-api-global-places-and-poi-data-service" target="_blank" rel="noopener nofollow">OpenPOI on RapidAPI</a>
<hr>
I’d love to hear your thoughts:
<ul>
<li>Feedback on the caching strategy?</li>
<li>Features you’d like to see next?</li>
</ul>
<hr>
#webdev #backend #opensource #geospatial #javascript #api

Murat

OpenPOI is a self-hosted alternative to Google Places API that uses OpenStreetMap data with a three-layer caching architecture (Redis, MongoDB, Overpass API). The system features a self-healing background worker that proactively fetches and caches POI data for surrounding areas using hexagonal scanning, ensuring millisecond response times for subsequent queries. It offers hybrid search capabilities, simplified opening hours parsing, and is available on RapidAPI with a free tier.

Tired of Google Places API Bills? Meet OpenPOI: A High-Performance, Self-Healing POI Service