CloudRegionCost answers two questions. Where is this one workload cheapest, once egress, a latency budget, compliance and carbon are counted: that is the ranking. And where should several workloads that talk to each other live so the total over a horizon is lowest, including what it costs to move: that is the placement optimiser.
Every number carries a date and a source. AWS and Azure compute prices are fetched from the providers' price APIs by a script in the pricing pipeline (live, as of 2026-10-10; live, as of 2026-10-10); GCP prices are read from Google's pricing pages (verified, as of 2026-10-10) until an API key is configured. Inter-region transfer rates for all three clouds, grid carbon intensity and latency estimates each have their own provenance, shown next to the results and on the methodology page. A scheduled job checks the data daily and opens a change list for review; the build refuses to publish live data older than 45 days. Unknown values are shown as unknown and never priced as zero.
What it will not tell you. Reserved and spot pricing, managed-service markups, support plans, licences, inter-AZ transfer and taxes are outside the model. The output is a shortlist with its reasoning, not a quote: confirm with the provider before you commit.
How it is built. A static site (Next.js, React, TypeScript) on S3 and CloudFront, deployed only by continuous integration through short-lived OIDC credentials, with automated tests for the prices, calculations and published pages. Everything computes in your browser; see privacy.
CloudRegionCost is independent and not affiliated with Amazon Web Services, Microsoft Azure or Google Cloud. Provider names are used to identify their services.