An AI-powered financial coach that helps traders make smarter financial decisions by providing behavioural insights based on their on-chain actions. Think of it as your personal "guardian angel" that guides you at crucial moments to keep you aligned with your financial goals.

Problem Statement

Traders and investors often create solid strategies but stray from them due to emotions, cognitive biases, and impulsive decisions—leading to unexpected portfolio risks. On top of that, crypto companies and influencers are constantly trying to pull you into chasing the next big thing.

Solution: Onchain Angels

We’ve developed an AI agent that monitors on-chain activity in real time and applies behavioural economics to portfolio shifts. After each trade, it analyses changes in risk exposure, detects potential cognitive biases, and sends personalised, psychology-backed questions—based on 23 well-researched behavioural patterns—via social media. This helps traders reflect, spot hidden biases, and make smarter decisions over time.

Ecosystem Overview 🌟

Each repository plays a vital role in ensuring the Onchain Angels solution works seamlessly together: 🤝

• 🤖 Agent: AI agent that processes data and generates personalised behavioural insights.
• 💻 Backend: Handles data processing and generation, integrating with external services such as Alchemy, Etherscan, CoinGecko, and OpenAI.
• 🌐 Frontend: Intuitive and lightweight user interface enabling profile creation and system interaction.
• 🔒 Nillion: Secure data management through Nillion Secret Vault, ensuring privacy and data integrity.
• 🐍 Nillion-sv-wrappers-py: Python utility library to communicate with Nillion db.

Repositories 🚀

Agent 🤖

Repository: https://github.com/onchain-angels/onchain-angels-agent Tech Stack: ElizaOS v0.1.9, TypeScript Deployment: Autonome

Houses the AI agent - a customized fork of elizaOS/eliza-starter, featuring custom character configurations, Farcaster and Twitter/X integration clients, and a plugin for retrieving users' latest transaction data from the backend.

Backend 💻

Repository: https://github.com/onchain-angels/onchain-angels-backend Tech Stack: Python, Django, Django REST Framework, PostgreSQL Deployment: Heroku https://api.onchain-angels.com/

This repository contains APIs for profile management and user activity webhooks. It handles all server-side logic, including wallet and token management, plus integrations with blockchain data providers (Alchemy, Etherscan, CoinGecko). Implements webhooks that summarize user actions and relay them to the agent.

Frontend 🌐

Repository: https://github.com/onchain-angels/onchain-angels-frontend Tech Stack: Next.js, Tailwind CSS, TypeScript Deployment: Vercel https://onchain-angels.com/

Web application where users can create profiles to receive AI-powered financial insights. User information is securely stored in the backend and utilized to generate personalized agent responses.

nillion-sv-wrappers-py 🐍

Repository: https://github.com/onchain-angels/nillion-sv-wrappers-py Tech Stack: Python, nilql

A Python library that provides wrappers for seamless integration with the Nillion Secret Vault. Inspired by NillionNetwork/nillion-sv-wrappers JavaScript implementation, it offers an abstraction layer for secure, decentralized data storage and encryption operations.

Nillion 🔒

Repository: https://github.com/onchain-angels/nillion Tech Stack: JavaScript, Node.js, nillion-sv-wrappers

Used for testing and schema development for Backend-Agent information exchange. Not implemented in the current submission but planned for future development