Fits whole user-chosen attempt-and-review blocks into explicit study-date minute budgets before a supplied exam date. Uses your priorities and repeat counts, keeps review time separate, and reports unplanned repetitions. It does not predict grades or mastery. Uses only adult user-controlled facts; review the result. No external lookup, storage or sending. Uses explicitly supplied structured inputs. Server computations are transient; the standalone website computes locally in your browser. Review the output before exporting it. No persistent account records, external source retrieval, sending, transactions or paid model calls. CSV/JSON review and export are available on the standalone website. MCP requests are capped at 128 KiB and computed responses at 256 KiB; individual field and record limits also apply.