Hostinger VPS pricing: check the refund clocks
Hostinger VPS pricing should not be treated as the cost of an unconditional free trial. The current refund policy adds a repeat-refund timing condition to the transaction window, and excludes VPS upgrades. Its policy and help summary describe different starting events for the repeat-refund clock. Keep transaction, prior request initiation and prior refund completion separate; a document-based timing screen cannot promise approval or a cash refund.
What the official documents actually establish
The English policy requires a VPS refund request within 30 days of the relevant transaction and, if there was another VPS refund, more than 180 days since its initiation. The help summary instead describes 180 days from completion of the previous refund and identifies the policy as authoritative. Both exclude KVM VPS upgrades. Those different event descriptions are preserved here rather than silently replaced with one date. Purchases made using cryptocurrencies are also excluded.
Our original document-clock register
We built a small evidence register with product, upgrade status, payment category, current transaction age, prior refund history and two prior-event ages. Each record carries an explicit unknown value when an event is missing. This avoids using a server creation date as an invoice date or using a completed payment event as the start of a prior refund request. The model checks whether two documented timing descriptions lead to the same screen result. It does not connect to a customer account or submit anything.
Reproduce the timing fixtures
All day values below are invented test inputs, not account observations. The rule thresholds 30 and 180 come from the linked official documents. No purchase or refund was tested.
| Case | Transaction age | Prior initiation age | Prior completion age | Result |
|---|---|---|---|---|
| A | 10 | not applicable | not applicable | screen_pass_not_approval |
| B | 10 | 181 | 179 | document_clock_disagreement_review |
| C | 10 | 180 | 180 | document_clock_disagreement_review |
| D | 10 | 181 | 181 | screen_pass_not_approval |
| E | 31 | not applicable | not applicable | outside_transaction_window |
| F | 30 | not applicable | not applicable | transaction_boundary_review |
| G | 10 | 181 | None | unresolved_prior_clocks |
| H | 10 | not applicable | not applicable | excluded_upgrade |
Download all inputs, thresholds, formulas and capture hashes ยท Run the original checker
Case A: no prior refund is an input, not a guess
Our first fixture selects a KVM VPS purchase, a non-crypto payment, no upgrade and a transaction age of ten days. It explicitly records no prior refund. The result is screen pass, not approval. That wording matters: it means only the limited known fields did not trigger a modeled rejection. It cannot establish service compliance, refund destination, paid amount or provider acceptance. In an actual private record, absence of a visible refund entry must not automatically become proof that no prior refund exists.
Case B: two dates can cross different sides of the threshold
We selected a prior request initiation age of 181 days and a completion age of 179 days, with the current transaction ten days old. Our legal-text timing flag is true because 181 is greater than 180. Our summary timing flag is false because 179 is less than 180. The output is document clock disagreement review. We keep both inputs and both comparisons visible instead of selecting the cheaper-looking interpretation. These elapsed ages are a synthetic pair, not dates taken from a buyer or a test refund.
Case C: the exact threshold is tested separately
For a second comparison fixture, both prior-event ages are set to 180. The literal greater-than comparison is false, while the summary-style at-least comparison is true. Our output again requests review. This test intentionally exposes the edge rather than using a rounded elapsed-month label. An actual event near a threshold needs exact timestamps and the applicable provider interpretation; our invented whole-day value does not decide the account outcome. The test is useful because changing either comparison operator would otherwise hide a meaningful implementation choice.
Case D and the two transaction-window controls
Case D moves both prior-event ages to 181 and both limited timing flags agree. It still returns screen pass, not approval. Case E changes the current transaction age to 31 and returns outside transaction window. Case F selects an exact age of 30 and returns transaction boundary review. We deliberately avoid promising a policy cutoff instant from a rounded day value. Together these fixtures separate the prior-history clock from the current transaction clock; extending one must not reset or repair the other in our register.
Cases G and H: missing history and a different purchase category
Case G keeps the selected prior initiation age at 181 but removes the completion age. The result stays unresolved prior clocks rather than inventing a completion date. Case H changes the upgrade field to true and receives excluded upgrade. The transaction date alone therefore cannot make the two modeled records equivalent. We also supplied a reversed chronology, where completion was older than initiation, and the checker rejected it. A different product was kept outside scope. These are input-integrity tests, not claims about a buyer eligibility.
Does passing this screen establish the money returned?
No. It does not calculate a refund amount or identify a destination. The help summary says a processed balance refund cannot be transferred to another payment source and that some methods need alternative procedures. Keep eligibility screening separate from cash recovery. We have not inspected an invoice, requested a refund or received money.
What remains unknown
This is a dated documentation comparison and a partial original evidence screen. It does not model every exclusion, account action, jurisdiction or provider decision. It supplies no live purchase price and does not predict a refund. Preserve the source version and exact private account events before drawing an account-specific conclusion.
Official evidence and test provenance
Captured October 12, 2026. All elapsed ages are project-authored synthetic inputs.