Welcome to Nofinity

This video shows how to transform Claude or ChatGPT into a fully-fledged financial advisor. The system includes a portfolio tracking dashboard, daily market analyses, and automated Telegram alerts. The key is not just good prompts, but a structured memory architecture that gives the AI continuous context about you and your finances.

The critical difference from simple chat queries is data-driven personalization. The AI is trained to behave like a McKinsey or Goldman Sachs analyst – based on real, personal data.

📁 Memory Folder Structure:

• Investor One-Pager: A questionnaire (6 phases) covering your financial situation, risk tolerance, goals, and mindset. The result is a summary saved as a .md file.
• Instructions.md: Defines the AI's behavior (e.g., "no fluff, direct, like a McKinsey advisor"). Includes a vital rule: "Read the one-pager and memory before every response. Update memory.md immediately when new information emerges – append, never overwrite, date every entry."
• Memory.md: A running log of all financial decisions, market assessments, and new insights. This document grows with every interaction, ensuring constant context retention.
• Financials Folder: Contains bank statements, profit & loss statements (e.g., monthly), so the AI knows your operating income situation.

Tip: Use voice prompting to complete the questionnaire in about 20 minutes. Create the folder on your desktop and connect it to Claude Code / Codex. This gives you full control over your data.

With the memory system in place, an interactive dashboard is created – no coding skills required, just instructions to Claude Code.

Dashboard Features:

• Live price data from Yahoo Finance for stocks and crypto
• Google Sheets integration – ideal for manual entry or API connections to brokers and crypto exchanges (e.g., CoinStats, Bybit, Binance)
• Portfolio overview including illiquid positions (e.g., VC investments)
• "Morning Brief" Generator: Automatically analyzes all positions in the context of current market events (e.g., Fed decisions, payroll data)
• Risk profile: Shows specific over- or underweight positions (e.g., "You're still 11% in NVIDIA – despite the trim")

Additional Analysis Options:

• Philosophy Check: Feed transcripts from finance experts (e.g., podcasts) and let the AI check how your portfolio aligns with their theses.
• Multi-Model Comparison: Use the same storage folder in ChatGPT Codex and Claude to get different opinions.

The biggest value-add: Daily, personalized market updates via Telegram – so you never miss a major development.

How to Set It Up:

1. Create a bot via BotFather on Telegram.
2. Copy the token into Claude Code.
3. Give the command: "Send me a market update every morning at 5:00 AM UTC based on the main headlines affecting my portfolio."

Message Content:

• Price movements over the last 24 hours
• Potential impact on the portfolio
• Concrete action suggestions
• Most important headlines overnight

Use cases: If Trump makes an announcement or the crypto market shifts quietly, you get an alert. For leveraged traders, overnight updates on open positions directly from the exchange are possible.

• Not financial advice: The AI sometimes hallucinates. All outputs must be double-checked.
• Does not fully replace professionals: The system is a "force multiplier" for better decisions – especially for passive investors who often react too late.
• Data sovereignty: The local folder belongs to you. Unlike cloud chats, you have full control over what the AI knows about you.
• Scalable: Beginners can use manual entry in Google Sheets. Advanced users can connect APIs (Bloomberg, Binance, brokers) for real-time data.

The core message: "AI is useless unless it knows everything about you." The memory system is the core of this transformation. It turns a generic chat assistant into a personal financial analyst who:

• knows your goals,
• understands your risk tolerance,
• is always up to date,
• informs you proactively – not just upon request.

Next steps: Miles announces a video on automated trade execution (e.g., a $20,000 AI trading challenge) and a Personal OS version of the dashboard.

Summary follows...

Summary follows...

Summary follows...

Summary follows...

Summary follows...

Summary follows...