Intrafere/MOTO-Autonomous-ASI

Employs a three-tier brainstorming-and-refinement pipeline where initial ideas feed into paper generation, completed papers form a second-tier brainstorm database, and final outputs emerge from a third tier that autonomously selects between short-form answers or book-length volumes (40,000+ words). Integrates with OpenRouter API and LM Studio for local or cloud-based model orchestration, with built-in embedding support via nomic-embed-text for offline RAG without external dependency costs. The reverse-order writing approach (body → conclusion → introduction) preserves creative freedom by avoiding early introduction constraints that would lock subsequent content.