CHORUS — Multi-Agent AI Trading Senate
Traditional trading bots are black boxes — one algorithm, no accountability, no transparency. CHORUS reimagines autonomous trading as a decentralized senate of specialized AI agents, each with a verifiable on-chain identity, a reputation score that evolves with every decision, and a cryptographically signed vote before any capital moves. Our system deploys five agents: a Trend Agent analyzing moving average crossovers, a Reversal Agent monitoring RSI oscillators, a Sentiment Agent scanning volume and price momentum, a Risk Sentinel with absolute veto power over drawdown limits, and a Meta-Agent that tallies weighted votes and executes trades through Kraken CLI. Every vote, risk check, trade intent, and outcome is logged permanently on the blockchain as a validation artifact — a tamper-proof receipt that anyone can audit. Agent reputation scores update automatically after each trade based on prediction accuracy, meaning agents that are consistently right gain more influence over future decisions. CHORUS targets the intersection of DeFi, autonomous AI agents, and verifiable trust — a space with massive growth potential as institutions demand accountability from AI-driven financial systems. Built on Base Sepolia with ERC-8004 identity standards, CHORUS is designed to be the trust layer that autonomous trading has always been missing.
Category tags:
.png&w=96&q=75){kind=small}
"1. Application of Technology: 4.5 / 5 Justification: Very strong execution of the prompt. Having the Python "meta-agent" actually query the ERC-8004 Reputation smart contract on Base Sepolia to fetch the historical score for each sub-agent before it tallies the vote is an incredibly smart way to use Web3 infrastructure in a Web2 Python application. 2. Presentation: 4.5 / 5 Justification: Fantastic branding and storytelling. The concept of an "AI Oracle Senate" where minority dissent is logged as a validation artifact is a tremendous narrative that immediately conveys the value proposition to institutions. The CHORUS.md document perfectly explains why this wins. 3. Business Value: 4.25 / 5 Justification: High value. In traditional finance, decisions are made by investment committees with different analysts presenting bull/bear cases. CHORUS digitizes this exact corporate governance structure using AI agents, making it a very familiar and attractive paradigm for institutional money managers looking to dabble in AI. 4. Originality: 4 / 5 Justification: The multi-agent approach is common, but creating a decentralized voting mechanism where the weight of the vote is dynamically determined by an immutable on-chain smart contract reputation score is a highly original marriage of AI and Crypto. ⚖️ Pros & Cons Pros: Meritocratic AI: Dynamically weighting the influence of each agent (Trend vs. Reversal) based on their actual, on-chain proven historical accuracy ensures the system naturally adapts to who is "right" in the current market regime. Minority Dissent Logging: By logging the votes of agents that disagreed with the final trade as a validation artifact on-chain, it creates a fascinating audit trail to review if a trade goes wrong (e.g., "The Risk Agent warned us, but the Trend and Sentiment agents outvoted it"). Excellent ERC-8004 Utilization: They didn't just use the blockchain to store receipts; they actively used the smart contract's Reputation state to drive the Python app's logic. Cons: Execution Latency: Tallying votes from 4 different AI agents, verifying their reputation via Web3 RPC calls to Base Sepolia, and then waiting for the meta-agent to execute via Kraken CLI means this architecture is strictly confined to slow, macro trading. It cannot do high-frequency or arbitrage trading. Basic Trading Alpha: The sub-agents themselves (Trend, Reversal, Sentiment) seem to rely on fairly standard Python/TA-Lib indicators rather than state-of-the-art predictive generative AI models. The innovation is entirely in the voting layer, not the prediction layer."
Sanem Avcil
For Innovators & Creators