The ultimate "One-Person On-Chain Hedge Fund" terminal. Sodex PowerOps leverages high-frequency SoSoValue data, institutional flows, macroeconomic indicators, and Google Gemini AI to orchestrate and execute autonomous trading agents on the SoDEX ValueChain.
Sodex PowerOps is an Agentic Finance Ecosystem developed for the SoSoValue x Akindo Buildathon (Wave 3, July 2026). By bridging the gap between raw research data and real-time execution, it allows retail traders to run automated, hedge-fund-level quantitative strategies on-chain.
For Wave 3, we have completely overhauled the platform's security, restructured the UI/UX for ultra-low latency, and developed a fully functional autonomous trading engine. By removing previous constraints and adding robust safety guardrails, Sodex PowerOps stands as a production-ready Web3 algorithmic suite.
Addressing critical feedback from the Wave 2 jury regarding client-side private key storage:
- Non-Custodial Connection: Users connect their EVM wallets via MetaMask or WalletConnect. All manual trading orders are signed using standard browser-level
eth_signTypedData_v4typed data signature relays. Raw private keys are never stored in the browser. - Session Key Delegation: For 24/7 background automation, users authorize a restricted Session Key on SoDEX with a user-specified expiry duration. The bots execute signed orders using this delegated session key, which is locked to trading functions only, ensuring the user's primary wallet remains completely secure.
A fully autonomous 24/7 self-operating trading platform run by an automated consensus of 4 specialized AI agents working in collaboration:
- Macro Agent: Evaluates SoSoValue ETF net flows, Dollar Index (DXY) stability, and SSI Sentiment indicators to vote on macro liquidity direction.
- Technical Agent: Computes live indicators (RSI, EMA cross, MACD, Bollinger Bands, and ATR volatility spreads) to vote on mathematical trends.
- Sentiment Agent: Leverages Google Gemini 2.0 to scan and score real-time news sentiment and headlines to vote on emotional market consensus.
- Risk Officer Agent: Evaluates margin health, current drawdown limits, and flash-crash triggers to approve, scale, or veto execution sizing.
- Consensus Voting & Auto-Rotation: The engine aggregates votes on every tick. If consensus is reached, it automatically deploys, closes, or rotates sub-bots (Grid Bot for High Volatility, DCA Bot for Trending Up, Signal Bot for Trending Down, or holds capital safely in stable assets during Consolidation).
- Live Discussion Feed: Displays a real-time conversational chat stream in the UI, showing the step-by-step rationales, scores, and votes of each agent.
- Centralized Risk Shield (Flash Crash Protection): If the Risk Officer detects a sudden price drop exceeding 3% in a 10s window, it overrides the consensus, liquidates all active positions, cancels outstanding orders, elevates the risk level to
CRITICAL, and halts operations until conditions stabilize.
The Signal Bot has been completely refactored from scratch into a full-screen Visual Quant Studio, introducing highly detailed technical, macro, and sentiment parameters:
- Multi-Signal Evaluation Engine: Fuses 9+ built-in indicators across distinct categories:
- Technical Oscillators: RSI, MACD, Stochastic RSI, Bollinger Bands, EMA Crossover, and Volume Spike.
- SoSoValue AI Sentiment: Evaluates live news sentiment scores fetched dynamically from SoSoValue hot news feeds.
- ETF Net Flows: Triggers signals based on daily net inflows/outflows of Spot Bitcoin and Ethereum ETFs.
- Macro Correlation: Monotors DXY (U.S. Dollar Index) movements to predict macro capital shifts.
- Custom Code Expression Engine: Quant developers can write custom JavaScript expressions (using variables like
rsi(14),ema(9),close,volume) to define their own math-based execution criteria. - Interactive Validation Sandbox: Allows users to run real-time sanity tests on their combined indicators against live chart markers before deploying live capital.
- Conflict Resolution Policies: Enforces strict risk management when signals conflict, offering policies like
IGNORE,CLOSE_ONLY(closes opposing trades), orCLOSE_AND_REVERSE.
We eliminated single points of failure related to API limits or key issues:
- Local AI Fallback: If the external Gemini API Key is missing or rate-limited, our custom deterministic quantitative engine automatically takes over.
- Seamless Diagnostics: It runs regime classification, calculates optimal bot configuration parameters, and performs wallet diagnostics locally without any downtime.
A completely new feature allowing users to mirror the trades of top-performing wallets on SoDEX:
- Neural Wallet Diagnostics: Inputting any EVM address triggers an AI-driven behavioral audit. It analyzes historical fills to measure win-rates, average hold times, max leverage, and categorizes their style (Conservative, Moderate, Aggressive).
- AI Co-Pilot Auditing: When the source wallet takes a trade, the backend copilot engine audits the order using Gemini 2.0 Flash against latest SoSoValue hot news sentiment. It can auto-reject or queue trades for manual approval if the risk score exceeds user-set parameters.
All automated bots now undergo a compulsory "Pre-Flight Check" prior to deployment:
- AI Risk Override Limit: Pauses bot execution if real-time sentiment risk exceeds a specific score.
- AI Trailing Stop Loss: Dynamically adjusts the bot's stop-loss price based on local ATR volatility metrics to lock in profits as market prices move in the trade's favor.
- Fee Drag Protection: Automatically restricts orders where estimated transaction fees (SoDEX 0.08% roundtrip) would eat up the projected profit.
- Bot Auto-Kill (Max Loss Switch): Shuts down the bot instantly if losses hit a user-configured limit (in USD or % equity).
- Flash-Crash Slippage Limit: Restricts execution during liquidity dry-ups.
Simulate quantitative strategies before deploying live capital:
- Mainnet Alignment: Pulls historical klines (real Binance data or volatile simulated klines) to run historical simulations.
- Fee Modeling: Demonstrates the impact of fee friction, proving why short-term high-frequency trades often underperform compared to macro timeframe expectancies.
- Performance Metrics: Computes real-time Sharpe ratio, profit factors, monthly expectancies, and accumulated volume.
- Sleek Navigation: Replaced the heavy sidebar navigation with a unified, professional floating Header Dock matching modern trading terminals.
- Ticker Tape Marquee: Shows live asset pricing, SSI sentiment indicators, and portfolio VaR.
- Real-Time Network Latency Indicator: Displays connection roundtrip time directly in the header dock with color-coded status alerts (Green/Yellow/Red) to inform traders of network performance.
- React Error Boundary: Robust fallback rendering catches unexpected runtime errors gracefully, showing clear debugging info and a reload option instead of crashing into a blank screen.
- Bloat Reduction: Cleaned up 20,000+ lines of legacy code, unused packages, and inactive features (like Telegram integration and news bot templates) to guarantee sub-second page loads.
The ultimate goal of the SoSoValue Buildathon is to enable developers to build a "One-Person On-Chain Hedge Fund." Sodex PowerOps fulfills this thesis across all three core dimensions:
- Macro & Quantitative Research (The Brain): A hedge fund starts with data. By fusing SoSoValue's institutional flows, Spot ETF inflows, SSI Sentiment indicators, and corporate treasury indexes with technical oscillators, the platform provides a complete market intelligence dashboard. The Neural Wallet Diagnostics tool acts as a research analyst, evaluating historical wallet performance and profiling risk.
- Collaborative Multi-Agent Consensus (The Committee): Instead of relying on basic, static rule-triggers, Sodex PowerOps replicates an investment committee. The Macro, Technical, Sentiment, and Risk agents debate and vote on every price tick. No execution happens unless the Risk Officer approves the consensus, ensuring institutional-grade decision logic.
- Algorithmic Non-Custodial Execution (The Execution Desk): A fund must manage risk and execute efficiently on-chain. Through Session Key Delegation, users delegate temporary trading rights to the autonomous bots while keeping their assets safely in MetaMask or WalletConnect. The central Flash Crash Risk Shield and Fee Drag Protection act as automated risk controllers, protecting capital from liquidity dry-ups and transaction fee erosion.
By packaging research, consensus modeling, risk management, and secure on-chain execution into a single unified terminal, Sodex PowerOps turns any individual trader into a self-sufficient, institutional-grade hedge fund.
- Frontend: React 19, TypeScript, Vite 8, Zustand v5 (Persisted slices).
- Styling: TailwindCSS v4 & Vanilla CSS (Modern dark glassmorphism).
- Execution Backend: Node.js, Express, WebSocket (Proxy client for secure API interactions).
- AI Engine: Google Gemini 2.0 Flash API (Audit copilot) + Local AI Engine.
- Data Feeds: SoSoValue API (News, ETFs, Sector Spotlight, Indices, Treasuries).
- Smart Wallet Signing: Ethers.js v6 + Web3 Browser Provider (
eth_signTypedData_v4).
- Node.js (v18+)
- npm or yarn
git clone https://github.com/keoyle52/SoDexTerminal.git
cd SoDexTerminal/sodexpoweropsCreate a .env file in the root of sodexpowerops and sodexpowerops/backend:
VITE_API_BASE_URL=http://localhost:5000
GEMINI_API_KEY=your_gemini_api_key
SOSOVALUE_API_KEY=your_sosovalue_api_keycd backend
npm install
npm run devcd ..
npm install
npm run devOpen your browser and navigate to http://localhost:5173.
- User Value (30%): Validates performance metrics (Sharpe, Drawdown) to protect capital while enabling users to run a "one-person hedge fund."
- Functionality (25%): Fully functional on-chain EIP-712 order execution with comprehensive backtesting and Wallet connection.
- Product Design (20%): Fully responsive modern dark-mode Glassmorphic UI with micro-animations and clean charts.
- Data/API Integration (15%): Ingests news, ETF, macro, fundraising, SSI index, and treasury data from SoSoValue alongside Gemini AI reasoning.
- UX & Clarity (10%): Clear consensus visual flows and explanatory cards on quantitative trading mechanics (e.g., fee friction).
This software is developed for educational and hackathon purposes. Trading cryptocurrencies involves significant financial risk. The authors are not responsible for any financial losses.