He Built a 12-Node n8n Workflow to Analyze One Stock. ChatGPT Plus Does It in One Prompt.

The Pipeline

Three HTTP nodes fetch 1-minute, 15-minute, and 1-hour candlestick data from 12data API. A fourth node pulls news articles from NewsAPI. Everything gets merged, aggregated, cleaned with custom JavaScript, and fed through two separate AI nodes — one for sentiment analysis, one for the final trade recommendation. Twelve nodes total for one stock ticker.

The Title vs. The Demo

The video is called 'I Built a Profitable AI Agent Day Trader.' The actual demo output? Technical recommendation: Hold. Entry price: Not available. Stop-loss: Not available. Target: Not available. The title says profitable. The demo says sit on your hands.

Three API Keys for One Answer

12data API for candlestick data. NewsAPI for news articles. OpenAI API for the AI models. Three separate accounts, three API keys, three rate limits to manage. 12data gives you 800 calls per month free — that's about 27 analyses per day if you only check one stock. Most day traders watch dozens.

What One Prompt Can Do

ChatGPT Plus with Code Interpreter can ingest a CSV of candlestick data, calculate RSI, MACD, moving averages, and trend lines, cross-reference with recent news it already knows about, and give you a reasoned recommendation — with the math shown. One prompt. No API keys. No merge nodes. No JavaScript cleanup.

Perplexity AI handles the news sentiment side in real-time with cited sources. Instead of piping NewsAPI through a separate sentiment analysis node, you ask Perplexity 'what's the current sentiment on AAPL' and get a sourced answer in seconds.

The Uncomfortable Truth

The custom JavaScript node in the workflow exists because the data pipeline is poorly structured — the AI model can't understand its own input without preprocessing. Purpose-built tools don't have this problem because they were designed to ingest financial data natively. The 25 minutes spent building this workflow is 25 minutes of not actually trading.