Drasko

Drakos is an AI-driven platform that helps users make smarter crypto investment decisions by analyzing the sentiment surrounding tokens across major social platforms. It uses autonomous AI agents that fetch real-time data — tweets, discussions, and mentions — from sources such as Twitter (X), Discord, and Telegram. These agents apply advanced natural language processing and sentiment-classification models to interpret public mood, identifying whether the overall sentiment around a token is bullish, neutral, or bearish. By translating raw community discussions into clear, data-driven insights, Drakos allows users to gauge market psychology before acting, reducing emotional bias and improving timing in their trades.

Beyond analytics, Drakos introduces an innovative layer of agent autonomy and customization. Users can deploy their own AI agents as on-chain smart contracts, giving them independent operational logic and transparent decision-making. Each agent can be configured to choose its own data sources, defining its base layer — whether it relies on Twitter for fast-moving trends, Discord for developer sentiment, or Telegram for community pulse. This decentralized model ensures flexibility and trust: every user’s agent becomes a self-governing entity that can evolve, learn, and operate according to the preferences of its creator.

In essence, Drakos merges AI intelligence with blockchain autonomy, enabling a network of user-owned agents that continuously scan the social web for sentiment signals. These insights empower investors to not just follow market movements but to understand the emotions and narratives driving them, ultimately transforming how decision-making happens in the decentralized world.

We built Drakos by combining the robustness of on-chain smart contracts with the adaptability of modern AI frameworks. On the blockchain side, we used Foundry for contract development, testing, and deployment. It allowed us to write modular, gas-efficient contracts that define the lifecycle and logic of each AI agent — from its creation and configuration to how it fetches or interprets sentiment data. Foundry’s scripting environment made it seamless to automate deployments and interact with agents directly from the command line during rapid iteration.

For the frontend, we used React with JavaScript to build an interactive and visually engaging dashboard. The React app serves as the user interface where individuals can deploy, manage, and monitor their on-chain agents. It communicates with the contracts through ethers.js, handling wallet connections, contract interactions, and real-time updates of sentiment scores. The frontend also integrates APIs that collect social data — tweets, Discord messages, and Telegram chatter — which are then passed through the AI pipeline for processing.

The heart of the AI logic runs on Eliza OS, which we used to orchestrate agent behavior and sentiment analysis. Eliza helped us give each agent reasoning capabilities — allowing it to fetch data dynamically, parse sentiment contextually, and even decide when to trigger certain alerts or recommendations. By linking Eliza’s AI orchestration with Foundry-deployed agents, we created a hybrid system where on-chain autonomy meets off-chain intelligence.

We also did a few hacky but creative things to make everything work smoothly. For instance, we built a custom bridge that lets on-chain agents communicate with off-chain Eliza processes via event listeners and webhooks — effectively turning contract events into AI prompts. This approach allowed our agents to remain decentralized while still tapping into real-world data streams. Overall, Drakos became a cohesive ecosystem where smart contracts, AI agents, and web interfaces work in tandem — each piece amplifying the other’s capabilities.