MarketForge
MarketForge is an AI-powered market intelligence platform that uses reinforcement learning to train smart agents to make optimal trading and negotiation decisions.
Core trick
A multi-agent LLM market runs continuous double auctions and natural-language bilateral negotiations, rewarding trades in an RL loop.
Description
MarketForge is an AI-powered market intelligence platform that uses reinforcement learning to train smart agents to make optimal trading and negotiation decisions. Created with OpenEnv, it functions as a virtual marketplace or "flight simulator" for traders where AI agents practice and master strategies safely before deployment.The platform trains Large Language Models (LLMs) through four core pillars of multi-agent interaction:Cooperation: Building supply chains to create compound goods.Competition: Participating in continuous double auctions for scarce resources.Negotiation: Proposing and accepting natural language bilateral deals.Coalition: Forming dynamic buying and selling alliances based on reputation.By running thousands of simulated trades and learning from a reward system , the AI agents help businesses eliminate pricing guesswork , navigate market disruptions , and accelerate deal-making.
README excerpt
--- title: Multi-Agent MarketForge emoji: 🏪 colorFrom: green colorTo: blue sdk: docker pinned: false app_port: 7860 suggested_hardware: t4-small tags: - openenv - multi-agent - market-simulation - grpo - reinforcement-learning --- # Multi-Agent MarketForge A **multi-commodity market simulation environment** purpose-built for training LLM agents in **cooperation**, **competition**, **negotiation**, and **coalition formation**. Designed as an OpenEnv v0.2.1 environment for the OpenEnv Hackathon. ## The Four Pillars of Multi-Agent Interaction | Pillar | Mechanism | Agent Behavior | |--------|-----------|---------------| | **Cooperation** | Supply chain formation | Agents combine raw commodities into compound goods (Bread = Wheat + Oil) | | **Competition** | Continuous Double Auction | Agents bid/ask on limit order book for scarce resources | | **Negotiation** | Natural language bilateral deals | Agents propose and counter-propose trade terms | | **Coalition** | Dynamic alliance formation | Agents form buying/selling groups with reputation tracking | ## Environment Features - **4 raw commodities** (Wheat, Iron, Timber, Oil) and **3 compound goods** (Bread, Tools, Furniture) - **8 heterogeneous agents** across 4 roles (Producer, Consumer, Trader, Speculator) - **12 stochastic market events** (droughts, embargoes, surpluses, festivals) - **Partial observability** -- agents only see own inventory + top-of-book prices - **Mixed-motive payoffs** with configurable self-