Modern neural networks have long ceased to be just "chatterboxes." With proper integration with trading terminals, they can analyze charts, generate indicator code, conduct strategy backtesting, and even review your trade journal. However, no prompt will print money for you. The AI's task is to remove the routine, but you still make the decisions. I have compiled 10 working prompts for Claude that transform it from a simple chatbot into a full-fledged trading assistant.
1. Chart Reading: Facts Without Forecasts
This prompt forces the neural network to connect to your TradingView via an MCP connector and describe the current market situation solely based on data: market structure, key levels, indicator readings. No "I think" — only "the price made a higher high." Requires TradingView MCP setup.
2. Indicator Generation: From Idea to Code in Seconds
You don't need to know Pine Script v6. Simply describe in words the condition under which the signal should trigger (e.g., "fast EMA 9 crosses slow EMA 21 from below with increased volume"). Claude will write ready-made code with visual markup and honestly warn you in which cases this indicator may give false signals.
3. Strategy Formalization: Turning "Buy on a Pullback" into Precise Rules
Everyone has a vague idea. This prompt forces the neural network to break it down into specific, measurable rules: exact entry price, exit conditions, stop-loss placement, and position size. The rules must be so clear that two different traders would trade them absolutely identically.
4. Backtesting: Verification on Historical Data
Before trading with real money, your strategy script is run through history. Claude writes the strategy code in Pine Script, accounting for a commission (0.1%), and outputs all trades, drawdown, and percentage of profitable ones. At the end, it provides criteria for distinguishing a truly good result from "curve fitting."
5. Loss Analysis: Your Personal Auditor
The most valuable prompt in the collection. You give the neural network a table of your trades, and it finds common patterns in losing ones: time of day, entry type, volatility. Then it suggests one specific filter to cut out such trades and honestly assesses whether this filter is "curve fitting" to past data.
6. Market Scanner: Selection by a Given Pattern
Instead of manually scanning dozens of coins, you give Claude a list of tickers and describe the desired situation (e.g., "price returned above the 200 EMA with increasing volume"). The neural network, via TradingView MCP, scans all assets and outputs only those that fully meet the condition, sorting them from best to worst.
7. Alert Setup: Meaningful Notifications
Claude sets up an alert on the chart that triggers at a precise event. The main feature is the notification text. You won't get an empty "signal triggered," but a specific message: which coin, at what level, what it means, and what action to consider.
8. Multi-Timeframe Analysis: Top-Down View
The same trade can look attractive on a small timeframe but go against the global trend. The prompt forces the neural network to check the consistency of weekly, daily, and 4-hour charts and issue a verdict: "ENTER," "WAIT," or "SKIP."
9. Journal Review: A Strict Coach on Your Trades
You upload the last 30 trades. The neural network acts as a strict coach: finds the most profitable technique, the most harmful habit, and also a technique you think works but actually doesn't bring money. At the end — one clear rule for your trading plan, backed by numbers.
10. Risk and Trade Size Calculation: Discipline Above All
The final prompt is about capital management. You enter account size, acceptable risk per trade (e.g., 1%), entry, stop, and take profit prices. The neural network calculates the exact position size, profit-to-risk ratio (R:R), and if R:R is below 1.5, it directly orders you to skip the trade.
Expert Opinion: Prompts 1, 6, and 7 require TradingView MCP connection — this is a key point for serious analytics. The rest work in a regular chat. Remember: a neural network is a powerful tool for removing routine and finding patterns, but the final responsibility for the decision lies with you. Do not blindly follow signals — use AI to test your hypotheses.