Aubm AI Agent Orchestration

Aubm uses a FastAPI backend, React frontend, Supabase for persistence and authentication, and multiple LLM provider integrations. It also includes budget tracking, retry flows, PDF export, marketplace-style agent templates, and optional vector memory when embeddings are configured. The target users are managers, team leads, department heads, and middle-management operators who need to coordinate work, evaluate AI-generated outputs, and turn strategic goals into traceable execution plans. The goal is to make AI agents usable for real operational workflows where planning, accountability, review, and evidence matter. The platform includes guided and expert modes, project dashboards, task approval flows, agent management, debate workflows, evidence intelligence, monitoring, team workspaces, audit logs, and final report generation. It is designed for use cases where AI output needs traceability rather than a single opaque answer. Each task can produce structured output, claims, evidence, and logs so that users can inspect how work was completed before approving it. Aubm uses a FastAPI backend, React frontend, Supabase for persistence and authentication, and multiple LLM provider integrations. It also includes budget tracking, retry flows, PDF export, marketplace-style agent templates, and optional vector memory when embeddings are configured. The goal is to make AI agents usable for real operational workflows where planning, accountability, review, and evidence matter.