A few days ago I started an experiment.
The question was pretty simple:
Could I build a useful, functional website that solved a real problem — without actually having a web server?
The problem I picked was one I've run into repeatedly: finding useful GIS API endpoints.
So I started building
TocsinData.io.
No traditional web hosting. No Apache or Nginx server sitting somewhere serving the site. No VPS behind it.
The site and API gateway run at the edge using Cloudflare Workers, with the project maintained through GitHub.
And then the experiment got a little out of hand.
It stopped being just a directory of GIS APIs.
TocsinData.io is becoming a gateway to useful public data from multiple sources, organized behind predictable endpoints.
Right now that includes data and services covering things like:
• USGS earthquake data
• Canadian geographic names
• FCC Census and geographic lookups
• NOAA coastal and ocean station metadata
• NASA natural-event categories
• National Weather Service weather and alert data
• World Bank country and reference data
• NHTSA vehicle recalls
• GBIF species and taxonomy matching
• FEMA disaster declarations
• OpenAlex research/publication metadata
• UK police force and crime-category reference data
• USAspending government reference data
• Federal Register agency information
• UK bank holidays
And that's only the beginning.
GIS data now has its own namespace, while non-GIS public APIs have their own home as well.
The goal isn't to blindly proxy every free API on the Internet. Each source has to be reviewed, bounded and handled according to its own technical and usage requirements.
What started as:
"Can I make a useful GIS API website without buying hosting?"
has turned into:
"How much useful public data can I make easier for humans, software and AI systems to discover and use — while still not having a conventional web server?"
Apparently quite a lot.
TocsinData.io
The experiment is still running.