CloudOverview

CloudOverview / Reference

Methodology & sources

Every estimate starts with a defined product, a dated rate card and a visible set of assumptions.

What we verify

We read official provider pricing and product documentation. The launch dataset was checked on 7 September 2026 and is versioned as 2026-09-07.1. A verification date means the linked documentation was checked; it does not mean we tested a provider, negotiated a quote or inspected its infrastructure.

Rates and factual qualifications are maintained together in structured data. The public pricing JSON includes source URLs, verification dates, product scopes, regions, capability states and price components. Profiles show the relevant source trail. The regional inventory is intentionally a subset rather than a claim of exhaustive coverage.

One engine, two presentations

The browser and the static worked examples use the same calculation implementation. Static totals are generated at build time. Interactive changes run locally in the browser; no pricing API or database is needed. URL parameters store the quantities and allowance setting. They do not freeze future prices; keep the dataset version with any saved estimate.

Totals are not ranked. Unknown required components produce an incomplete estimate, never a zero-priced component. A documented zero, such as an included standard request class, is distinct from a missing price. Preview support and unknown capabilities stay separate from supported features.

Units, currency and time

All user inputs are decimal GB: one GB is 1,000,000,000 bytes. AWS and Wasabi use binary billing units in this model, so the engine converts to GiB (1,073,741,824 bytes) before pricing. Wasabi’s per-TB rate is divided by 1,024 to obtain a GiB rate. R2 and B2 use decimal units. Bunny and Scaleway are interpreted as decimal GB; the byte definition was not explicit on their checked pricing pages, so this is an exposed assumption.

USD and EUR stay separate. There is no hidden exchange-rate conversion. All estimates exclude tax, discounts, account credits, premium support and unmodeled services. Capacity is held at the entered billing-period average. Scaleway uses a constant 730-hour month, with capacity rounded up to a whole GB. The other rows use their published monthly rates; this is not a calendar-accurate byte-hour invoice reconstruction.

Allowances and tier rules

R2 includes Standard monthly allowances and rounds paid storage to GB and paid requests to million-operation units. S3 uses tiered Standard storage and internet delivery; the shared 100 GiB internet allowance is assumed available when enabled. B2 applies its free capacity allowance and the contractual egress-to-storage ratio. Standard request calls are included at the current checked B2 rate.

Wasabi applies its monthly minimum and withholds estimates for shorter-than-90-day retention or traffic beyond the published egress guideline. It does not estimate object-age-dependent deleted storage. Bunny includes one Standard HDD region, EU/NA Standard CDN delivery and an account-minimum top-up, assuming this is the only workload in the account. Scaleway includes its outbound allowance only when available.

The allowance switch removes shared monthly free allowances. It does not remove contractual egress ratios, minimum spend, billing rounding or acceptable-use conditions. Different providers aggregate allowances differently, so adjust the setting based on the account that will actually pay.

Limits of the comparison

Only hot/standard storage is in scope. No negotiated rates, cold retrieval, minimum-size cold objects, replication add-ons, compute, image processing, monitoring, KMS or migration fees are modeled. The calculator accepts up to one million GB for each byte input and ten billion operations per request class. Large scenarios may require a commercial quote and service-limit review even if arithmetic is available.

The products have different delivery and resilience models. Equivalent stored bytes do not establish equivalent durability, latency or throughput. No original provider benchmarks have been performed for this release.

Updates and corrections

A maintainer checks the official sources, updates the versioned dataset, validates the schema, regenerates examples and runs tests before publishing. A failed research or refresh attempt leaves the last verified dataset intact. We do not change the verified date merely because a build ran.

Use the correction route to prepare a source-backed report. Include the exact product, region, currency and effective date so a change can be evaluated without guessing.

Official source index