Automated daily swing‑trade ideas from end‑of‑day (EOD) data, scored by an LLM, logged to Google Sheets, and pushed to Telegram.

What this workflow does

• Fetches EOD quotes for a chosen stock universe (example: NSE‑100 via RapidAPI).
• Cleans & filters the universe using simple technical/quality gates (e.g., price/volume sanity, avoid illiquid names).
• Packages market context and feeds it to OpenAI with a strict JSON schema to produce top swing‑trade recommendations (entry, target, stop, rationale).
• Splits structured output into rows and logs them to a Google Sheet for tracking.
• Sends an alert with the day’s trade ideas to Telegram (channel or DM).

Ideal for

• Retail traders who want a daily, hands‑off idea generator.
• PMs/engineers prototyping LLM‑assisted quant sidekicks.
• Creators who publish daily trade notes to their audience.

Tech stack

• n8n (orchestration)
• RapidAPI (EOD quotes; pluggable data source)
• OpenAI (LLM for idea generation)
• Google Sheets (logging & performance tracker)
• Telegram (alerts)

Prerequisites

1. RapidAPI key with access to an EOD quotes endpoint for your exchange.
2. OpenAI API key.
3. Google account with a Sheet named Trade_Recommendations_Tracker (or update the node).
4. Telegram bot token (via @BotFather) and destination chat ID.

You can replace any of the above vendors with equivalents (e.g., Alpha Vantage, Twelve Data, Polygon, etc.). Only the HTTP Request + Format nodes need tweaks.

Environment variables

Google Sheet schema

Create a Sheet (tab: EOD_Ideas) with the headers:

Date, Symbol, Direction, Entry, Target, StopLoss, Confidence, Reason, SourceModel, UniverseTag

Node map (name → purpose)

1. Trigger – Daily Market Close → Fires daily after market close (e.g., 4:15 PM IST).
2. Prepare Stock List (NSE 100) → Provides stock symbols to analyze (static list or from a Sheet/API).
3. Fetch EOD Data (RapidAPI) → Gets EOD data for all symbols in one or batched calls.
4. Format EOD Data → Normalizes API response to a clean array (symbol, close, high, low, volume, etc.).
5. Filter Valid Stock Data → Drops illiquid/invalid rows (e.g., volume > 200k, close > 50).
6. Build LLM Prompt Input → Creates compact market context & JSON instructions for the model.
7. Generate Swing Trade Ideas (OpenAI) → Returns strict JSON with top ideas.
8. Split JSON Output (Trade‑wise) → Explodes the JSON array into individual items.
9. Log Trade to Google Sheet → Appends each idea as a row.
10. Send Trade Alert to Telegram → Publishes a concise summary to Telegram.