ApplianceAPI for AI Builders & Autonomous Agents

Why AI Agents Need Structured Appliance Data

LLMs cannot scrape retailer sites or access proprietary databases at runtime. They need structured APIs that return clean, typed data in a single call.

Appliance data is domain-specific and not well-represented in training data. Model numbers, serial number formats, and recall databases require specialized knowledge that general-purpose models lack.

ApplianceAPI bridges this gap: every endpoint returns structured JSON with human-readable fields, confidence scores, and consistent formatting that agents can parse, summarize, and act on.

Example Prompts

• "Use ApplianceAPI to look up model WRS325SDHZ and return the photo URL and product title."
• "Decode serial number FH212345X for a Whirlpool washer. What year was it manufactured?"
• "Check if GE model GFE28GYNFS has any active recalls."
• "Write a function that takes a brand and model, calls ApplianceAPI, and returns a formatted summary."
• "I have a service ticket for a Samsung WF45T6000AW. Get the product photo and specs from ApplianceAPI."

Why LLMs Need Structured Appliance Data

Large language models do not have reliable access to appliance-specific data. When asked "what year was this Whirlpool made?" or "is this GE model recalled?", an LLM has to guess or hallucinate. ApplianceAPI provides the structured, verified data that makes these queries answerable.

Key gaps that ApplianceAPI fills for AI systems:

• Serial number formats — each brand uses a different encoding scheme (55+ brands, 12 algorithm families)
• Model-to-brand mapping — WRS325SDHZ is a Whirlpool, GFE28GYNFS is a GE, but LLMs cannot reliably make this connection
• Recall status — CPSC data changes frequently and is not in training data
• Product photos — LLMs cannot retrieve images by model number without an API

Example Agent Workflows

Here are concrete ways AI agents use ApplianceAPI:

// Repair Estimate Agent
// 1. Customer provides model number
// 2. Agent calls ApplianceAPI for specs + photo + age
// 3. Agent estimates repair cost based on age and type

const res = await fetch(
  'https://www.applianceapi.com/api/v1/images?brand=Whirlpool&model=WRS325SDHZ',
  { headers: { Authorization: 'Bearer YOUR_KEY' } }
);
const data = await res.json();
// data.title = "Whirlpool 25 Cu. Ft. Side-by-Side Refrigerator"
// data.confidence = "high"
// data.imageUrl = "https://..."

// Inspection Assistant (Claude Code)
// Prompt: "Decode serial CR3748566 for a Whirlpool washer.
//          What year was it manufactured? Is it past typical lifespan?"

// Agent calls: POST /appliance-date/decode
// Body: { brand: "whirlpool", serial: "CR3748566", model: "WTW5000DW" }
// Response: { year: 2012, month: "Sep", confidence: "high" }
// Agent: "Manufactured September 2012. At 13+ years old,
//         this exceeds the typical 10-14 year washer lifespan."

Supported AI Platforms

ApplianceAPI works with any platform that can make HTTP requests:

• Claude Code — direct HTTP calls via curl or fetch. See Claude Code examples →
• ChatGPT — custom GPT actions, function calling. See ChatGPT integration guide →
• Cursor / Windsurf / Copilot — any AI coding tool with HTTP access
• LangChain / LlamaIndex — as a tool in agent chains
• Make / Zapier / n8n — no-code automation. See no-code examples →

What AI Agents Cannot Do Without an API

LLMs are trained on static data with a knowledge cutoff. They cannot reliably answer questions about:

• Current recall status — CPSC issues new recalls regularly. Training data is months old.
• Serial number decoding — Each brand uses a proprietary format. 55+ brands, 22 algorithm families. LLMs cannot reliably decode these.
• Verified product photos — LLMs do not have image retrieval capability by model number.
• Real-time model verification — Is this model number valid? Is it in our database? Only a live API can answer.

ApplianceAPI provides the structured, verified, real-time data that fills these gaps.