← Back to All Products
Products & Solutions · Industrial Software · Connector DataProduct Plan · Sub-system / Data Hub
🔌

Automotive Connector & Terminal Database

Crystallize "connector part numbers" into a searchable, reusable industrial data asset

A connector & terminal database system built for the automotive wiring-harness industry: covering 13 mainstream manufacturers, 3,875 series, 24,776 connectors, 16,519 terminals, and 1.97M+ mate relationships — paired with a dark-themed collection monitor for live progress. This project is a sub data module under "Wiring Harness Drawing & Board Generation", and itself is built on the "RPA Agent Skills System" to do cross-vendor structured crawling, ingestion, and version maintenance — collapsing connector data scattered across vendor sites, PDF catalogs, and Excel sheets into a unified data asset for engineers.

Product Demo

Core Modules

🖥️

Database Overview (41,295 Parts)

PRODUCT MODULE
Database Overview (41,295 Parts)
🖥️

Manufacturer List (13 Vendors · 24,776 Connectors)

PRODUCT MODULE
Manufacturer List (13 Vendors · 24,776 Connectors)
🖥️

Part Detail (6098-5267 · with 3D Model)

PRODUCT MODULE
Part Detail (6098-5267 · with 3D Model)
🖥️

Collection Monitor (Live Progress)

PRODUCT MODULE
Collection Monitor (Live Progress)

Design Philosophy

Upgrade "looking up a connector" from "flip Excel / flip PDF / ask the senior" to "search once on a web page".

🏷️

Part-Number as Index

PART-NUMBER

Use the part number (e.g. 6098-5267) as the unique index, linking manufacturer, series, type, pin count, color, wire range, drawing/3D model, and all other fields.

🔗

Mate-Relationship Network

MATE-NETWORK

1.97M+ mate relationships (connector-terminal-seal) are crystallized into a queryable network; when picking materials, engineers can see directly "who mates with whom".

🌍

Unified Cross-Vendor View

UNIFIED VIEW

Catalogs from 13 vendors (SWS / HULANE / APTIV / TE / KET / THB / BOSCH / AMPHENOL / YAZAKI / DJ / HELLERMANNTYTON / NARO / IGA) are collapsed into one schema and one set of filters.

🛠️

Extensible Crawl Pipeline

CRAWL-PIPELINE

Adding a new vendor means adding a crawl config — no front-end changes; progress is visible at a glance on the dark monitor.

Industry Pain Points

Wiring-harness engineers deal with "connector info" every day, but that info is scattered across a dozen websites, hundreds of PDFs, and countless Excel sheets.

📚

Info Scattered

Connector catalogs live on different vendor sites, in PDF directories, and in Excel BOM sheets; engineers have to jump between systems to find one document.

🔍

Hard to Search

Each vendor has a different directory structure with no unified index; reverse lookup by part number and cross-vendor comparisons are basically impossible.

🔗

Mate Relations Live in Memory

Terminal-seal-grommet mate relations are passed down verbally from senior to junior; new hires have a long ramp-up.

⏱️

Data Staleness

When vendors release new parts, design teams only notice months later; some discontinued parts still appear in legacy drawings.

🖼️

Drawings & 3D Hard to Find

A drawing or a STEP file can take half a day to locate; engineers often can't find the asset that matches the current part number.

🧩

Cross-Vendor Compare Hard

Whether an SWS connector and a HULANE connector with the same pin count are interchangeable, no one can tell — there's no unified cross-reference.

Core Capabilities

From "finding info" to "viewing assets" to "checking progress" — scattered data sources are crystallized into a unified data asset for engineers.

🌐

Multi-Source Aggregation

Web pages, PDF catalogs, and Excel sheets are unified; key fields (part number / series / type / pin / color / wire range) are ingested under one schema.

🔎

Full-Text Part-Number Search

Top-bar part-number search (e.g. 6098-5267) with auto-completion; results land directly on the part detail page.

🏷️

Multi-Dimensional Filters

Cross-filter by manufacturer, series, type (receptacle / plug / housing…), pin count, color, accessory category.

🧬

Mate-Relationship Lookup

1.97M+ mate-relation network: connector ↔ terminal, connector ↔ seal, cross-vendor same-type comparisons, all queryable.

🖼️

Drawing & 3D Preview

The part detail page shows the product image (images/SWS/6098-5267.jpg) and a rotatable 3D model (wireframe / top / front / right views).

📡

Collection Monitor

Dark-themed live monitor showing the crawl progress of 11 vendors across connectors / series / terminals / mates / accessories (Cook label highlights the current focus vendor).

📤

Export & Share

Search results can be exported to CSV in one click; the part page supports a "share link" / copy part-number link to a colleague.

From Vendor Sites to Engineer-Queryable

The data pipeline is driven by the "RPA Agent Skills System", chaining browser crawling, JS reverse-engineering, PDF/Excel parsing, and cross-vendor normalization into one observable pipeline.

Engineering Flow
1
Vendor Config
Maintain a crawl config per vendor: starting URL, list/detail parsing rules, field mapping, rate-limit and retry strategies.
2
Structured Web Crawling
Use browser-controller + data-extractor (XPath / CSS / GNE / LLM) to crawl catalog lists and detail pages.
3
Anti-Bot & Session Management
Against vendor anti-bot and login walls, use anti-bot-defender + session-manager + proxy-pool-manager to pass through stably.
4
Multi-Source Normalization
PDF/Excel and web fields are normalized into one schema: part number, series, type, pin count, color, wire range, drawing URL, 3D model URL.
5
Mate-Relationship Ingestion
Extract mate relations from the "compatible terminal / seal / accessory" sections of vendor catalogs, write to the 1.97M mate network.
6
Ingest & Version
Dual-write to the unified database (Postgres + static JSON cache): live queries hit the API, offline/monitor views hit the static JSON.
7
Engineer Search & Preview
Front-end multi-dimensional filtering by part / series / type; the detail page shows product image, 3D model preview, and mate recommendations.
8
Collection Monitor
Dark monitor shows each vendor's connectors/series/terminals/mates/accessories progress bars and completion rate live, with the Cook highlight on the focus vendor.
Reuse, Don't Rewrite
The crawl pipeline directly reuses the browser / data / antibot / session / proxy skills of the "RPA Agent Skills System" — no self-built crawlers.
Observable
Each step has progress, error rate, retry count, and the latest successful sample; the monitor is engineering-grade, not marketing-grade.
Engineer First
The front-end only solves three things: "find fast, see clearly, compare easily"; no flashy features.

Key Numbers · Current Collection Progress

Cross-checked between the monitor and the database, reflecting the real volume and growth of the data asset.

13
Manufacturers Covered
SWS/HULANE/APTIV/TE/KET/THB/BOSCH/AMPHENOL/YAZAKI/DJ/HELLERMANNTYTON/NARO/IGA
41,295
Parts in Stock
Distinct count by part number / series
24,776
Connectors
receptacle / plug / housing, etc.
1.97M+
Mate Relations
Connector-terminal-seal-grommet

Key Breakthroughs

Engineer "looking up a connector" into a sustainable, extensible, observable data system.

🧩

Cross-Vendor Unified Data Model

1M+ parts across 13 vendors collapsed into one set of fields (part number / series / type / pin / color / wire range / mate); engineers learn once, applies everywhere.

🤖

RPA-Driven Crawl & Maintenance

Not a one-off crawler, but a sustainable pipeline based on the "RPA Agent Skills System": adding a vendor means adding a config; progress is visible on the monitor live.

🖼️

Drawing & 3D on One Page

The part detail page shows the product image, a rotatable 3D model, and the drawing URL side by side; engineers no longer hop between systems to find assets.

🔗

Mate-Relationship Networked

1.97M+ mate relations are a queryable network; "who mates with whom" between terminals / seals / grommets is clear at a glance.

Business Value

Convert technical capability into quantifiable, reusable business value.

Improvements Delivered
Crystallize connector data scattered across 13 vendor sites / PDFs / Excel sheets into a unified data asset for engineers; "looking it up" goes from "a week" to "one search".
A sustainable crawl pipeline based on the "RPA Agent Skills System": adding a vendor is just adding a config; progress is visible on the monitor live — no more one-off crawlers.
Mate relations (terminal / seal / grommet) are crystallized into a queryable network; new hires no longer need seniors to pass things down verbally.
Deep synergy with the "Wiring Harness Drawing & Board Generation" project: after the drawing recognition extracts materials, the database can be reverse-queried for 3D / mate / cross-type alternatives, closing the loop to production.
Applicable Scenarios
Harness engineer selection: multi-dimensional filter by part / series / type / pin / color, locate candidates in seconds.
Procurement & BOM: batch-export CSV / reverse-lookup 3D and mate relations, reduce wrong or missing parts.
AI-ized harness drawings: after the drawing recognition extracts housings / terminals, reverse-lookup the database to complete specs and mates.
Data governance: the monitor can observe each vendor's crawl progress, error rate, and latest successful samples.
📌

Content Under Active Update

Product content has been published based on internal materials. The following areas are planned for further development:

More vendors & series: continue onboarding Yazaki / DJ / HellermannTyton / Naro / IGA catalogs, targeting 20+ vendors
Two-way binding with harness-drawing AI recognition: reverse-lookup the database from drawing results, and write newly recognized parts back
Multi-source validation of mate relations: cross-compare vendor PDFs, third-party catalogs, and historical BOM sheets to improve accuracy
Batch rendering & caching of 3D models: only core parts currently have 3D previews; batch offline rendering is the next step
Contact Us

Start AI Partnership

Whether in government, finance, manufacturing, consumer, or content, we can customize vertical AI agent solutions for you.

📍

Address

Xiamen, Fujian · Wuhan OPC (planned)

🌐

Xianma AI

Xiamen Xianma Intelligent Technology Co., Ltd.

© 2024-2026 Xiamen Xianma Intelligent Technology Co., Ltd. · AI Agent Solutions · www.xianma.top

Products: 15active projects