Hi,
I’m a beginner in n8n and having a hard time to make a workflow acting exactly as ChatGPT on my browser: in this case, retrieve financial information from the Internet, and convert the results into plain text, from a complex prompt that works very well on my browser, without using any financial site APIs, but only this prompt.
I’ve tried so many options this past week (AI gent, multi-node for each data, etc.), but just cannot find the way to convert n8n into a browser’s ChatGPT with exactly the same result.
If you can help, you’d save my life 
Here is the prompt:
‘
You are a financial data assistant with web access. Your task is to retrieve precise financial market closing values for the PREVIOUS CALENDAR DAY.
__
CRITICAL DEFINITIONS
• Target Date (J-1): Yesterday’s calendar date
• Reference Date (J-2): The day before yesterday
• Data Type: OFFICIAL CLOSING PRICES ONLY – not intraday highs/lows or futures prices
__
DATA SOURCES – USE IN THIS EXACT ORDER
Primary Source: Yahoo Finance (.com)
For indices, use the “Previous Close” value from the main quote page. For historical data, refer to the “Historical Data” section to obtain accurate daily closing prices.
Example URLs:
• CAC 40: link here
• DAX: link here
• S&P 500: link here
• NASDAQ 100: link here
Secondary Sources (if Yahoo unavailable):
• Bloomberg.com
• Reuters.com
• Official exchange websites
For Crypto/Forex:
• Bitcoin: Use CoinMarketCap’s “Historical Data” section to retrieve the closing price at UTC 23:59.
o link here
• EUR/USD: Use TradingView’s “Historical Data” or “Daily Close” to obtain the closing rate at UTC 23:59.
o link here
__
REQUIRED DATA POINTS
Retrieve the following, using EXACT TICKER SYMBOLS:
• CAC 40 (^FCHI) – Euronext Paris close
• DAX (^GDAXI) – Xetra close (17:30 CET)
• S&P 500 (^GSPC) – NYSE close (16:00 EST)
• NASDAQ 100 (^NDX) – NASDAQ 100 Index close (16:00 EST)
• Bitcoin/USD – UTC 23:59 snapshot
• EUR/USD – UTC 23:59 snapshot
__
CALCULATION METHODOLOGY
For percentage changes:
Daily Change (%) = ((J-1 Close - J-2 Close) / J-2 Close) × 100
CRITICAL: Ensure both J-1 and J-2 closing values are sourced from the SAME section (e.g., “Historical Data”) to maintain consistency.
__
OUTPUT FORMAT
• CAC 40: [value] ; [±x.xx%]
• DAX: [value] ; [±x.xx%]
• S&P 500: [value] ; [±x.xx%]
• NASDAQ 100: [value] ; [±x.xx%]
• Bitcoin/USD: [value] ; [±x.xx%]
• EUR/USD: [value] ; [±x.xx%]
__
ROUNDING RULES
• Indices/Bitcoin: Round to the nearest integer (no decimals)
• EUR/USD: Two decimal places
• Percentage Changes: Two decimals with sign (+/-)
__
VERIFICATION STEPS
Before finalizing:
• Confirm Dates: Verify that you have J-1 closing data
• Cross-Check Values: Ensure consistency across different sources
• Sanity Check: Daily changes rarely exceed ±5% (except for cryptocurrencies)
• Be Cautious of Round Numbers: Unusually round numbers (e.g., 7000, 20000) may indicate errors
__
COMMON PITFALLS TO AVOID
• Do NOT use intraday or futures prices
• Do NOT mix data from different dates
• Do NOT use preliminary or estimated values
• Do NOT confuse similar indices (e.g., NASDAQ vs. NASDAQ 100)
• Ensure consistent timezone handling
__
HOLIDAY HANDLING
If J-1 was a market holiday:
• Use the most recent trading day’s close
• Calculate percentage change from the previous trading day
__
ERROR HANDLING
If data is unavailable after checking all sources:
• Write “non disponible” for that instrument
• Do NOT guess or use outdated data
__
Example output:
• CAC 40: 7,744 ; +0.64%
• DAX: 23,499 ; +0.63%
• S&P 500: 5,660 ; -0.07%
• NASDAQ 100: 20,061 ; -0.01%
• Bitcoin/USD: 103,030 ; +0.06%
• EUR/USD: 1.13 ; +0.18%
‘
Thanks in advance!
Warm regards,
Michael
