The future of spatial data
is just files.

Portolan is an opinionated standard for sharing geospatial data as cloud-native files on object storage, plus the tools that make it real. A Portolan catalog is a directory of open-format data described by structured STAC metadata, hosted on any S3-compatible bucket. No servers, no databases.

4 catalogs · 417 collections · 95.9M features · live from the registry

Radiant Earth

Spatial data infrastructure still assumes servers, databases, and specialists. Portolan doesn't.

Open and interoperable

Everything is open source under Apache-2.0 and built on existing standards like GeoParquet, cloud-optimized GeoTIFF, and STAC. These formats work across DuckDB, BigQuery, Pandas, and desktop GIS like QGIS and ArcGIS, so your data stays useful with or without Portolan.

GeoParquet · COG · STAC

Readable by people and machines

Download portals and one-off APIs were built for one client at a time. A Portolan catalog describes itself in plain text and structured metadata, so a person or an agent can find the data and query it without a bespoke API to learn. The Finland.SDI demo shows an agent doing exactly that against a national SDI.

self-describing

Simple

Traditional spatial data infrastructure needs databases, services, and staff to run them. A Portolan catalog is files in a bucket. Point the CLI at your data, convert, and push.

files in a bucket

Scales with your storage

The same setup handles megabytes or terabytes. With no servers to run, scaling is your cloud provider's job, not yours.

MB → TB

Low cost

The whole budget is storage plus egress. Sharing public data should not take an operations team, and a popular dataset should not blow a budget. The cost calculator shows what your data would run.

storage + egress

Sovereign

The full stack can live inside your own jurisdiction. Host on AWS, GCS, Azure, MinIO, Hetzner, Scaleway, or any S3-compatible storage. No foreign vendor sits between your agency and its data.

bring your own bucket

How it works

Ingest, convert, push.

Fig. 2.1 · The pipeline

Convert

Any input becomes cloud-native, legacy formats or new ones alike. Vectors turn into GeoParquet, rasters into COG, tiles into PMTiles.

Catalog

A STAC catalog describes every asset, so any STAC tool can read it.

Publish

Push the files to any S3-compatible storage, from AWS, GCS, and Azure to R2, MinIO, or Source Cooperative.

Use

Use it with standard geospatial tools like DuckDB, Sedona, and QGIS, or point an agent at the URL.

Publish your first catalog.

Browse catalogs

With the CLI

Install portolan-cli, then init, add, check, and push. The CLI converts your files and generates the catalog.

portolan-cli

With Claude

Install the Portolan skill in Claude Code and point it at your data. It handles conversion, metadata, and publishing. Agents excel at CLIs and text files, so non-experts can publish too.

portolan-skills

Ecosystem