HiveForge: Digital AI Workforce Platform

Overview HiveForge is a mission-based, multi-agent orchestration platform designed to automate complex, multi-step business goals. Traditional LLM chats fail when faced with layered, cross-functional operations. HiveForge solves this by decomposing human objectives into a structured Directed Acyclic Graph (DAG) of dependencies and executing them through a specialized digital workforce. How It Works Coordinated Planning: The AI Coordinator analyzes a goal (e.g., launching a store) and generates a structured, database-persisted task dependency graph. Parallel Scheduling: Tasks are scheduled dynamically via a dual-engine queue (using BullMQ on Redis for scalability, with a zero-setup in-memory fallback). Specialized Workers: Independent agents (Research, Finance, Marketing, Operations) execute tasks in isolation, passing data down the pipeline. Live Command Center: A Next.js dashboard visualizes the execution graph with glowing state borders and animated SVG connection lines representing data flow, accompanied by real-time token and cost telemetry. Technical Stack & Architecture Built as a modular monorepo using Nx and pnpm, the backend features NestJS API modules, Winston telemetry correlation IDs, and Prisma database transactions. The frontend is powered by Next.js and Tailwind CSS v4. HiveForge supports multiple AI providers (Fireworks AI, OpenAI) and includes a context-aware mock provider for offline demonstrations.