Sarkoa vs Self-Hosting Tile Infrastructure

Teams land here from different places. Some already run GeoServer, Tegola, Martin, or a custom COG tiler and wonder whether a hosted product is worth it. Others have a Shapefile on a laptop and just need a stakeholder link by Friday. Both are valid; they optimize for different outcomes.

This comparison is about time to a usable shared map and people-hours, not a claim that self-hosting is wrong.

What you are really buying

“Tile infrastructure” sounds like a serving problem. In practice the job is usually:

1. Get data in a web-friendly shape
2. Put it somewhere reliable
3. Let the right people open it (link, embed, desktop GIS)
4. Keep it updated without a weekend fire drill
5. Capture feedback on the map itself

If your product is a tile platform, owning that stack can be strategic. If your product is the dataset, the analysis, or the decision, the stack is overhead.

What self-hosting usually owns

A realistic DIY path often includes ingest scripts (GDAL, tippecanoe), object storage and lifecycle, a tile server or Worker, CDN and cache purge rules, auth for public vs private, rebuild pipelines when sources change, and monitoring so a misconfigured bucket does not burn egress. None of that is impossible. Together it is a product you maintain.

What Sarkoa owns for you

Sarkoa’s loop is upload → optimize → view → share → review:

• Accept common geospatial uploads and keep the source files
• Build viewer-ready copies so pan and zoom stay fast
• Serve maps and tiles without you standing up a tiler
• Share via public map links and website embeds
• Open layers from the QGIS plugin or CLI, or copy tile URLs for external clients
• Collaborate with comments and live annotations on the map

You still choose visibility, styling, and who gets the link. You do not provision the conversion fleet or the tile edge. Details live on pricing and the FAQ.

Where self-hosting still wins

• Hard isolation: air-gapped or VPC-only rules that forbid a SaaS data plane
• Exotic pipelines: custom projections, unusual formats, or pre-processing that must stay in-house
• Mature platform already paid for: the marginal cost of one more layer is near zero
• Maximum control: you must dictate every cache header and binary dependency

Where hosted wins for map outcomes

• Time-to-share: file to stakeholder link in minutes (Hosting Your First Map)
• Client delivery: embed or share without teaching viewers how to install anything
• Desktop continuity: same assets in the browser and in QGIS
• Review in place: annotations and comments instead of screenshot threads
• Fewer moving parts: no weekend pages for conversion jobs or tiler deploys
• Predictable product surface: storage and tile usage on a plan instead of surprise egress

A simple decision rule

Ask: Is operating a tile platform a competitive advantage for us, or a tax on shipping maps people can act on?

If it is a tax, start hosted and revisit when compliance or scale forces a private plane. If it is an advantage, self-host, and measure success by the same outcomes: time to share, who can open the map, and how feedback comes back.

Try the outcome, not the checklist

Spin up a free Sarkoa workspace at app.sarkoa.com, upload a sample layer, share a map, and optionally open it in QGIS. Separately, time your current DIY path for the same file through to a link a non-technical person can open. The gap is usually the answer, more than any feature matrix.

---

This post was written with AI assistance and reviewed for accuracy. Capabilities and pricing change over time; verify against the live product and pricing pages before procurement decisions.