NBA salary cap · 2023 CBA · 2026-27 league year
A front-office console for NBA roster construction. A pure TypeScript rules engine checks every proposed trade against salary matching, both aprons, hard-cap triggers, cash limits, roster bounds and the Stepien rule, then explains the verdict with the arithmetic.
2026-27 team salary against the five lines. August is when contenders duck the second apron.
◆ SAC $201.5M · $1.1M over the tax
⌇ scale $130.0M to $235.0M · ◆ SAC is the home desk · tap a team for its cap sheet · data as of 2026-08-04
60 seconds
Three modules
Build a two-team deal from real rosters, cash and picks. The Rule Ledger itemizes every CBA check, pass or fail, with the arithmetic written out. Every trade is a shareable URL.
Team salary as a gauge against the five lines, dollar distance to each, three years of commitments, and which exceptions its salary level allows, including the hard cap each one would trigger.
Contract-aware player cards with 2025-26 percentiles computed against every qualified player in the league, plus a radar to compare up to four.
Under the hood
/engine is pure TypeScript: no React, no Next.js, no I/O. Each CBA rule is a function you can audit in one sitting, and boundaries are tested to the dollar ($209,015,000 is not over the first apron; $209,015,001 is).
engine/tradeRules.tsRosters are Zod-validated JSON snapshots with a source and access date. A pre-build gate re-sums every team through the engine and fails the deploy if a single dollar disagrees with the published total.
scripts/validate-data.tsThe UI reads the same REST routes you can curl. POST /api/trade/validate takes a proposal and returns the full ledger, so the trade logic is reusable outside this interface.
app/api/