LLM Chains vs Single Calls in n8n: My Caption Generation Experiment

Source: Dev.to

The Two Approaches

Approach A: 3‑Step Chain (Haiku)

Break the task into micro‑steps:

1. Extract key points from the data.
2. Draft structure using the Step 1 output.
3. Polish and add metadata (hashtags, CTA, etc.).

Approach B: Sonnet Single Call

Send the entire prompt to a more capable model in one go, letting it handle the whole flow.

Test Script (Outside n8n)

# Setup
npm init -y
npm i @anthropic-ai/sdk dotenv
export ANTHROPIC_API_KEY="your_key_here"

index.mjs

import "dotenv/config";
import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });

// Sample data structure
const image_details = [
  { order: 1, attributeA: "value1", attributeB: "value2" },
  { order: 2, attributeA: "value3", attributeB: "value4" },
];

async function callClaude({ model, system, prompt }) {
  const res = await client.messages.create({
    model,
    max_tokens: 1000,
    system,
    messages: [{ role: "user", content: prompt }],
  });
  return res.content?.text ?? "";
}

// Approach A: Haiku 3‑step chain
async function haikuChain() {
  const model = "claude-haiku-4-5";
  const system = "You are a caption editor. Be concise.";

  // Step 1: Extract data
  const step1 = await callClaude({
    model,
    system,
    prompt: `Extract key descriptions from this data:\n${JSON.stringify({ image_details }, null, 2)}`,
  });

  // Step 2: Create structure
  const step2 = await callClaude({
    model,
    system,
    prompt: `Based on this, create hook + body + CTA:\n${step1}`,
  });

  // Step 3: Add tags
  const step3 = await callClaude({
    model,
    system,
    prompt: `Add hashtags to this caption:\n${step2}`,
  });

  return { step1, step2, final: step3 };
}

// Approach B: Sonnet single call
async function sonnetSingle() {
  const model = "claude-sonnet-4-5";
  const system = "You are a caption editor. Be concise.";

  return await callClaude({
    model,
    system,
    prompt: `Create a complete caption (hook + body + bullets + CTA + tags) from:\n${JSON.stringify({ image_details }, null, 2)}`,
  });
}

// Run both
const resultA = await haikuChain();
const resultB = await sonnetSingle();

console.log("=== Haiku Chain ===\n", resultA.final);
console.log("\n=== Sonnet Single ===\n", resultB);

Preparing Data in n8n’s Code Node

// Wrong: LLM can't work with IDs alone
const image_details = transformedImages.map(img => ({ id: img.id }));

// Right: Provide the full content the LLM needs
const image_details = transformedImages.map((img, index) => ({
  order: index + 1,
  ...img, // Spread ALL attributes the LLM needs
}));

return [{ json: { image_details } }];

Root cause: Only image IDs were passed, leaving the model without any descriptive context.
Fix: Spread the entire object (...img) so the LLM receives all relevant attributes.

Metric Comparison

Verdict: For captions (or any creative writing), continuity of context outweighs modest cost savings.

Pattern & Use‑Case Guidance

Key takeaways:

• Context > Cost: Fragmented chains break narrative flow for creative tasks.
• Data quality matters: Supplying rich information is essential; chaining cannot compensate for missing context.
• Haiku’s niche: Excellent for speed and highly structured tasks, but Sonnet excels when “feel” matters.

References

• n8n Advanced AI Docs
• Claude Sonnet 4.5 Announcement
• Claude Haiku on AWS Bedrock