FOMO

Our project is an AI-powered portfolio management system that integrates 1inch Fusion+ to enable efficient, gas-optimized token swaps across chains. The system allows users to define a custom asset pool, choose or create an allocation algorithm, and evaluate its performance through a built-in backtest framework.

We incorporate a large language model (LLM) as a user-facing advisor to guide decision-making — helping users select strategies, explain swap impacts, and answer questions about portfolio health using natural language.

On the backend, we use 1inch's Developer Portal APIs to fetch wallet balances, token prices, and transaction data. These data sources enable seamless portfolio tracking and strategy automation.

Together, these components create a personalized, intelligent, and cross-chain-capable portfolio manager — all powered by 1inch’s infrastructure and designed for both casual and advanced users.

core mechanism for automated portfolio rebalancing. We directly call the Fusion+ smart contracts via the 1inch API and let it handle the pathfinding and settlement logic, which simplified our swap logic significantly.

Our backend is written in Python, structured as a modular microservice system. Key modules include:

LLM agent , used for guiding users through portfolio design and swap decisions. We use langchain with gpt4o model with RAG to save 1inch token info and use system prompts as well as user information to consult and guide user friendly and concisely.

Backtest engine, which takes historical price data (pulled via the 1inch token price API) to simulate user-defined allocation strategies.

For the frontend, we used TypeScript + React (TSX) ....

We also use the 1inch Developer Portal APIs extensively:

Token Price API to support both backtesting and real-time monitoring.