SOFTWARERADAR — CURSOR PLAN AND CREDIT REVIEW

Guide: https://softwareradar.co.uk/guides/cursor-legacy-plan/
Official public documentation checked 16 October 2026. Account-specific terms and notices need separate review. No Cursor account, subscription, payment, top-up, cap, code change, agent request or switch was operated.

plan-identification.csv has four fictional summaries: A has later legacy monthly pools; B has current weekly usage and prepaid credit; C has an older request-count arrangement; D says only Ultra, which is insufficient. These are editorial cases, not real account observations. Actual plan and renewal evidence remain blank.

credit-ledger.csv is an idealised USD individual-account example inside one unchanged monthly cap period. Opening prepaid credit is USD50; the assumed cap is USD15. Consume USD12: balance38, cap room3. Buy USD10 credit: balance48, cap room still3. Consume USD3: balance45, cap room0. A weekly included-pool reset leaves both values unchanged in this cap period. The assumed purchase is not a published top-up package or actual provider charge. No expiry or enforcement lag is modelled; real enforcement can briefly lag. Eligible included usage is separate from paid additional consumption.

Opening50 + new credit10 - paid usage15 = closing45. New cash spent on credit in these events is10, not25: buying credit and consuming it are different measures. The base subscription, opening credit purchase history, taxes, discounts, conversion, actual model token costs and complete invoice are outside this example. No real allowance or universally available request count is inferred.

account-review.csv contains six proposed evidence questions. Actual account/plan, renewal, result, invoice/usage evidence, reviewer and decision are intentionally blank. The files do not connect to an account, determine entitlements or authorise changing billing.