IKOS — Global OEM Auto Parts Search Platform

Case Study · July 2026

A global OEM auto parts search engine built entirely with AI. Search millions of parts across worldwide distributors, find cross-references, and integrate OEM search into your own platform via API.

Case Study | July 2026

ikosdb.com | RapidAPI

If you export auto parts, you know this problem: a buyer emails you about part number BJ-3-12. Your factory makes it. You know there are distributors in Germany, Poland, and Turkey who stock it. But finding them — actually finding their websites, their catalogs, their part listings — means hours of manual Google searching, opening 50 distributor sites, and Ctrl-F'ing part numbers one at a time.

IKOS solves this. It is a search engine, but instead of searching the entire web, it searches global auto parts distributor databases. Type in an OEM number. IKOS tells you who stocks it, where they are, and what else they carry.

---

What IKOS Does

Three things, each one solving a specific problem for auto parts exporters and buyers:

OEM number search. Enter any OEM or aftermarket part number. IKOS finds every distributor worldwide listing that part in their catalog. Not scraped from spam directories — pulled from actual distributor inventory databases.
Cross-reference discovery. A radiator might be listed as BJ-3-12 by the OEM, as RAD-312 by a German distributor, and as BJ312 by a Polish warehouse. IKOS cross-references these variations. You stop missing matches because of formatting differences.
Distributor profiling. For each match, IKOS shows the country, the distributor's brand focus, and the types of parts they carry. You know who you are reaching out to before you send the first email.

---

Why I Built It

I run jjradiator.com, a B2B auto parts export site with 15,000 products. Finding buyers used to mean:

Searching Google for "radiator distributor Europe" → generic results, mostly Alibaba and Amazon
Opening individual distributor websites → slow, each one has a different catalog structure
Manually checking if my exact part numbers appeared → hours per supplier

One afternoon I searched "BJ-3-12" on Google and got 8 pages of results. Two were actual distributors. The rest were dead links, Alibaba listings from competing manufacturers, and an expired eBay auction from 2024. That was the afternoon I decided to build a tool that searches the right data, not the whole web.

IKOS was born from that frustration. I could not be the only exporter spending more time searching for buyers than talking to them.

---

How It Works

IKOS runs on three layers of AI:

Targeted web collection. Instead of scraping the entire internet, IKOS focuses on known distributor catalogs, parts databases, and trade listings. The collection layer knows the difference between a distributor's inventory page and a spam blog.
Vector-based fuzzy matching. Part numbers are messy. BJ-3-12 and BJ312 are the same part. AA-123/456 and AA123-456 are the same part. IKOS uses vector search to match variations without requiring exact string matches.
LLM cross-referencing. When a distributor uses an internal code instead of the OEM number, IKOS uses language models to identify the match based on product descriptions, dimensions, and application data — not just the part number.

The technical approach is the same manager-worker pattern I use across all my AI projects: Claude handles the complex reasoning (strategy, cross-referencing logic, quality checks), while DeepSeek and specialized models handle the volume work (search execution, data extraction, batch processing).

---

Try It or Integrate It

IKOS is available in two ways:

Web search: Go to ikosdb.com and search any OEM number. Free to try. No sign-up required.
API access: For developers building procurement platforms, parts catalogs, or ERP integrations — the IKOS OEM Search API on RapidAPI gives you programmatic OEM search without building the scraping and matching layer yourself.

API Quick Reference

Endpoint: GET /api/search?q={OEM number or keyword}

Example request:

GET https://ikos-auto-parts-oem-search.p.rapidapi.com/api/search?q=8D0121251
X-RapidAPI-Key: your-key
X-RapidAPI-Host: ikos-auto-parts-oem-search.p.rapidapi.com

Example response:

{
  "query": "8D0121251",
  "count": 12,
  "results": [
    {
      "code": "RA-10029",
      "brand": "AUDI",
      "model": "A4 B5 (8D2) 2.4/2.8",
      "oem": "8D0121251BB",
      "core": "630×398×32",
      "confidence": 90
    }
    ...
  ],
  "is_paid": false
}

Free tier vs paid:

	Free	Paid
Search by OEM / keyword	✅	✅
Brand, model, core dimensions	✅	✅
Confidence score	✅	✅
Full OEM number	Partial	✅ Full
Results per query	Up to 50	Up to 50
Rate limit	Standard	Higher

Free tier is enough to test and evaluate. Paid unlocks full OEM numbers and higher rate limits for production use.

---

What IKOS Is Not

It is important to be clear about what the tool does and does not do:

IKOS is not a marketplace. It does not process orders or payments.
IKOS is not a lead list generator. It helps you find distributors faster, but you still research and contact them.
IKOS is not a pricing tool. It shows who carries a part, not what they sell it for.

The goal is narrow and deliberate: answer the question "who stocks this part number?" faster and more accurately than a manual Google search.

---

IKOS was built using the same AI workflows I document on this site:

How I Turned a 600-Page PDF Into a Product Database — the data pipeline that feeds IKOS searches
The Manager-Worker Pattern — how Claude runs strategy while DeepSeek handles volume
How to Automate B2B Lead Collection and Follow-up — what to do with the distributors IKOS finds

---

Questions about IKOS or the API? Send me a message — I reply personally.