What Is n8n?
n8n is an open-source workflow automation tool that connects apps, APIs, and AI models and no coding experience required.
In this use case, you’ll use n8n to:
Fetch real-time data from an exchange (like Binance)
Analyze it with an AI model
Send a buy or sell signal to your 3Commas DCA or Signal bots
What You’ll Need
A 3Commas account and at least one active bot (DCA or Signal Bot)
A n8n account (self-hosted or cloud)
An OpenRouter account to access AI chat models like DeepSeek, Claude, or GPT
Basic familiarity with APIs and webhooks
How It Works
Your workflow in n8n will look like a chain of nodes (steps).
Each node performs one task and passes its output to the next.
Here’s the overall flow:
Trigger Node – defines how often your workflow runs
HTTP Node – fetches market data from your chosen exchange
Aggregate Node – bundles and cleans that data
AI Agent Node – analyzes the data and decides what to do (buy/sell)
Code Node – interprets the AI output and formats it as JSON
IF Node – checks the AI’s decision and routes it to the correct action
HTTP Node (Webhook) – sends the final signal to your 3Commas bot
Step-by-Step Setup
Step 1: Trigger Node
Choose when and how often to start your workflow (e.g., every 2 hours or every 3 days).
Example: Use the “Schedule Trigger” node.
Step 2: HTTP Node – Fetch Market Data
Method:
GETExample Endpoint (Binance):
https://api.binance.com/api/v3/klines?symbol=RENDERUSDT&interval=1m&limit=100
This fetches the last 100 candles for the RENDER/USDT pair on a 1-minute timeframe.
You can replace these parameters with your own trading pair and timeframe.
Step 3: Aggregate Node – Prepare the Data
Combine the fetched data into a single bundle so it can be processed by the AI.
This step “cleans” the raw API output for the next node.
Step 4: AI Agent Node – Analyze the Market
Connect your OpenRouter API key here.
Choose your AI model (e.g., DeepSeek R1, Claude 3, or GPT-4).
Prompt example:
Analyze this market data and decide whether to BUY, SELL, or HOLD. Provide the result in JSON format like: {"signal": "BUY", "confidence": "0.82"}
Tip: You can also ask the AI to provide a short reason for its decision.
AI agent chat model
Once you setup your AI agent node, you need to choose an AI agent chat model, this is where you will need an open router account and connect it to your AI agent chat model via API keys.
Once the API keys are created then go to your n8n account and create new credentials.
Then you can choose which AI model you want to use. For this example, we can use the free version of DeepSeek R1 model.
Step 5: Code Node – Interpret AI Output
Use a simple script to clean and parse the AI’s output into a usable JSON.
Example snippet (JavaScript):
// Step 1: Get raw string from AI Agent
let raw = $json.output || "";
// Step 2: Clean unwanted markdown or formatting artifacts
raw = raw
.replace(/```json/g, '')
.replace(/```/g, '')
.replace(/^[\s\r\n]+|[\s\r\n]+$/g, ''); // trim whitespace
// Step 3: Try parsing as JSON
let result = {};
try {
result = JSON.parse(raw);
} catch (err) {
// Fallback: use regex to extract signal manually
const match = raw.match(/"signal"\s*:\s*"?(\b(?:buy|sell|no signal|nothing)\b)"?/i);
if (match) {
result.signal = match[1].toLowerCase();
} else {
result.signal = "unknown"; // Still not found
}
}
// Step 4: Return clean object
return [{ json: result }];
This ensures your output is clean and ready for logic checks.
Step 6: IF Node – Decision Logic
Set the condition:
If
signal = BUY→ send to BUY HTTP NodeElse → send to SELL HTTP Node
You can also expand this with a Switch Node for more complex logic (e.g., hold, rebalance, alert).
Step 7: HTTP Node – Send Signal to 3Commas
Method:
POSTPaste your 3Commas webhook URL (from your DCA or Signal bot).
Example payload:
{ "action": "buy", "pair": "RENDER_USDT", "source": "n8n_AI_Agent" }When triggered, this sends your AI-generated signal directly to your bot in real time.
DCA Bot:
Signal Bot:
Pro Tips
You can chain sub-workflows for risk filters, volume checks, or custom indicators.
Use multiple AI nodes to compare strategies or cross-verify signals.
Log AI decisions for performance tracking.
Example Use Case Summary
Step | Task | Example |
1 | Trigger | Run every 2 hours |
2 | HTTP Node | Get RENDER/USDT data from Binance |
3 | Aggregate | Combine candle data |
4 | AI Agent | DeepSeek R1 analyzes trend |
5 | Code | Clean AI output |
6 | IF Node | Route BUY or SELL |
7 | HTTP Node | Send to 3Commas bot via webhook |
Next Steps
This is a simple starting workflow. Once you’re comfortable, try extending it to:
Add stop-loss or profit-taking conditions
Integrate sentiment data (e.g., from Twitter or CoinGecko)
Include multi-exchange strategies for cross-validation
n8n’s flexibility allows you to scale from this basic setup into full AI-driven trading automation.
n8n’s flexibility allows you to scale this setup into fully AI-driven trading automation. For help with n8n configurations or node issues, please contact n8n’s official support team.