
TableKeeper is a restaurant reservation platform inspired by modern booking systems such as OpenTable. It allows users to discover restaurants, view restaurant details and available tables, create reservations, view and manage bookings, cancel reservations, and amend existing bookings. The project was developed using an AI Dark Factory workflow, where specialized coding-agent seats collaborate through planning, backend implementation, frontend development, and independent quality verification. The factory uses dedicated roles for architecture and coordination, backend engineering, frontend engineering, and quality review. The implementation focuses on reliable reservation behavior, authentication, restaurant and table availability, reservation ownership, idempotent operations, validation, atomic booking changes, and protection against conflicting reservations. Stage 1 was independently tested locally and subsequently verified using the official Tablekeeper shipped harness in isolated mode, achieving a Stage 1 PASS. The factory workflow also demonstrates defect detection, handoff between specialized agents, corrective implementation, and independent verification.
6 Oct 2026