Tetraaaa: Binary MLM Network Platform with Automated Placement, Level Credits & Direct Referral Tiers

Network marketing platforms require complex, rules-driven logic at every join event: where in the binary tree does the new member go? Who gets paid, and how much? How is the credit pool tracked to prevent over-distribution? How are direct referral bonus tiers computed without double-awarding across monthly cron cycles? Building this on a Node.js stack required a custom automated tree placement engine, a sale-pool-based credit sourcing system, an admin fallback for underfunded positions, and a high-water mark bonus system for direct referral package upgrades — all with Sequelize ORM and PostgreSQL.

We built Tetraaaa with a Node.js + Express backend using Sequelize ORM on PostgreSQL. The automated placement engine starts from the root of the network tree, scans for the first available position from left to right, skips any positions already at maximum depth, and places the new user instantly — recording their position with a full parent chain and purchase history. Immediately after placement, the upline reward engine walks up the parent chain, calling creditUserOrAdmin at each level: if the target user has a Sale record with (sale_amount - reward_distributed) >= amount, the amount is drawn from that pool; otherwise the admin is charged and an AdminRewardLog entry records the reason (sale_limit_exceeded or no_sale_row). The directReferralCron uses a mutex flag to prevent overlapping runs, seeds default packages on startup via findOrCreate, then processes all due users: for each, it fetches all direct referrals and their most recent sale amounts via a single DISTINCT ON batch query, evaluates against packages sorted ascending by tier, finds the highest qualifying package, computes the delta against referral_bonus_credits (the high-water mark), and adds only the incremental difference — preventing double-awarding on tier advances. Withdrawal requests deduct credits immediately at submission; rejection refunds credits back. Admin getDashboardTotals runs four parallel aggregate queries in a single response for the overview cards.

Backend: Node.js + Express + Sequelize ORM + PostgreSQL. JWT authentication (bcrypt hashed passwords). node-cron direct referral package job (configurable CRON env). Parallel admin aggregate queries for dashboard overview. DISTINCT ON SQL batch query for referral sale amounts. JSONB product images. ARRAY product IDs. ENUM user roles and withdrawal statuses. Frontend: Next.js 14 App Router + React + Tailwind CSS. Public storefront + admin dashboard + user wallet.