FSQ OS Places is free. So why pay for place search?

In November 2024, Foursquare open-sourced its global places dataset. FSQ OS Places (Foursquare Open Source Places) is over 100 million points of interest (POI), released under the Apache 2.0 license, free for commercial use, and refreshed monthly. That was a genuinely good move, and the geospatial community is better for it.

So the obvious question: if the data is free, why would anyone pay for a place search API?

Because the license cost going to zero doesn’t make the cost disappear. It moves. It moves out of a contract and into your serving stack: a search index, a server, uptime, and a monthly re-ingest you now own. The decision is who carries that weight.

When is free FSQ OS Places genuinely enough?

Free FSQ OS Places is enough when you can run a data pipeline and the places are an input to something you compute in batch: enrichment, analytics, a one-time geospatial join, a model’s training set. Connect to the Places Portal, Hugging Face, or Snowflake offering with the data tool you already use, query the release, done.

It stops being enough the moment you need a live query API: “what’s near this coordinate, right now, as JSON, every time a user loads the page.” The dataset doesn’t answer requests. You’d be building the part that does.

What FSQ OS Places actually gives you

A real, sizable, openly licensed dataset. Concretely:

• Over 100 million POIs across more than 200 countries and territories (Foursquare OS Places).
• 20+ core attributes per place: name, address, locality, region, country, latitude/longitude, website, social handles, and category, among others (Foursquare schema).
• Apache 2.0 license, with Foursquare’s launch announcement saying the data is available for commercial use (Foursquare announcement).
• Monthly refreshes, so the data keeps moving with the real world (release notes).
• Delivered through the Places Portal using an Iceberg catalog, with additional Hugging Face and Snowflake offerings listed in Foursquare’s access docs.

That’s a strong foundation. For a lot of batch and analytics work, it’s the whole answer, and it costs nothing.

What it doesn’t give you: a query API

The dataset is files, not a service. Search, infrastructure, uptime, and refreshes are on you. Here’s a fair sketch of what self-hosting a live search endpoint involves.

First, ingest and filter. Connect to the current release and load the slice you need somewhere you can query. Tools like DuckDB, Spark, and Athena are good fits for exploration and batch work. You’ll want to filter too: a global dump includes places you don’t need and quality you’ll want to threshold.

Then, build a search index. Proximity search (“nearest N places to this lat/lon, matching this query”) is not what the raw dataset does well. You stand up something that does: Postgres with PostGIS, Elasticsearch, or a vector or geo index of your choice. You design the schema, the ranking, and the relevance behavior.

Then, host it and keep it up. That’s a server, a deployment, monitoring, and an on-call expectation if anything depends on it. A search endpoint that’s down is a feature that’s down.

Then, do it again next month. Foursquare ships a new release monthly. To stay current you re-ingest, re-filter, and re-index each time, ideally without downtime, and reconcile whatever changed. This is the part people underestimate: self-hosting open data isn’t a project, it’s a subscription you pay in engineering time.

None of this is exotic. It’s all standard work with standard tools. The question is whether it’s your work.

The three real options

Here’s the decision laid out plainly.

Self-host when you have the appetite for a data pipeline, want total control over filtering and ranking, and the places feed something you compute rather than serve live. The dataset is free and good. If your team already runs this kind of infrastructure, this is the cheapest path by a wide margin.

Pay Foursquare when you need what their open dataset deliberately holds back. Their paid Pro and Premium tiers add attributes the open set doesn’t carry: ratings, photos, tips, hours, popularity, amenities, and richer chain data (Foursquare pricing, Pro vs Premium schema). Current public pricing shows Pro endpoints with 500 free calls, then $15.00 per 1,000 up to 100,000 calls, stepping down at higher volume; Premium endpoints start at $18.75 per 1,000. If venue richness is your product, or you specifically want Foursquare’s data and global coverage, their hosted API is the right answer. Verify current rates against their pricing page before you commit.

Use a hosted open-data API when you want live proximity search without owning the stack, and you don’t need the rich-attribute layer. That’s where we fit.

Where Good Enough Maps fits

Good Enough Maps is hosted proximity place search over open Overture Maps Foundation data. One GET request with a server-side bearer key returns the places near a coordinate as JSON: names, coordinates, categories, addresses, and distances. We currently serve 39 million places across 18 countries and territories, refreshed as Overture releases pass validation.

There’s a direct connection to the topic here. Overture’s places theme draws from multiple sources, and in its September 2025 release Overture added about 6 million POIs from FSQ OS Places (Overture release notes). So Foursquare’s open data partially flows into what we serve, alongside contributions from Meta, Microsoft, and others (Overture places sources).

Pricing is flat with hard caps and no overages: $0 for 10,000 searches a month, $19 for 100,000, $99 for 1,000,000. When you hit the cap, the API returns a 402 and stops. It never bills you for the overage.

Now the scope. We’re narrower than both alternatives, on purpose. This isn’t autocomplete, it isn’t geocoding, and it isn’t a maps platform. It does one thing: find places near a point. If you need address autocomplete, map tiles, or the rich venue attributes Foursquare sells, we’re the wrong tool, and we’d rather you knew that now.

For a side-by-side, see Good Enough Maps vs the Foursquare Places API. To see the search return real results, try the live demo on the homepage without signing up.

FAQ

Is the Foursquare Places API free?

Partly. Foursquare offers a free open dataset (FSQ OS Places) under Apache 2.0, and their hosted Pro API endpoints include 500 free calls under the June 2026 pricing update. Beyond that, Pro calls bill at $15.00 per 1,000 up to 100,000 calls, less at higher volume, and Premium fields like photos, tips, and ratings bill from the first call (Foursquare pricing).

Can I use FSQ OS Places commercially?

Yes. The dataset is released under the Apache 2.0 license, and Foursquare’s launch announcement describes it as available for commercial use (Foursquare announcement). You’re responsible for hosting and serving it yourself.

Is Foursquare data in Overture?

Yes. Overture’s September 2025 release added about 6 million POIs from FSQ OS Places to the places theme, licensed under Apache 2.0 (Overture release notes). Foursquare sits alongside Meta, Microsoft, and other sources in Overture’s places data (Overture places sources).

Do I have to choose just one of these?

No. Self-hosting pairs fine with a hosted API as a fallback, and a hosted proximity search pairs fine with Foursquare’s paid attributes if you need both breadth and depth. The three options aren’t mutually exclusive; they’re three places the serving cost can land.

The data being free is real, and it’s a good thing. What’s left to decide is who runs the part that turns files into answers.

Last validated 2026-06-12.