A beam of light watches the warp yarns on high-speed looms for the workshop; two AIs — one checks production for the shift lead, the other watches raw material prices for the boss. We built this textile MES the hard way, one step at a time.
In the past two years when textile mills talk MES, two fears stand out: one is the "showpiece system" that lies flat after purchase — dashboards look pretty, but nobody in the workshop uses it; the other is a feature list piled sky-high, but on production scheduling, work reporting and traceability, you ask three questions and get zero answers.
After building systems long enough, we increasingly believe one thing: **the first tier isn't about having many features, it's about every point solving a real problem and running for real.

One, first solve the hardest pain point at the hardware level

Every textile mill knows: what do water-jet and air-jet looms fear most? Warp breaks. The traditional drop-wire approach — some too sensitive, some miss — quietly produces defect fabric and downtime.
We didn't take the old road of "add an alarm in software"; instead we redid a drop-wire-free optical beam warp-break detection device from the root — using optics to judge warp yarn state in real time. We compressed 10 invention patents** here, grinding from main control board, power supply to firmware piece by piece.
Pretty software is no match for one piece of equipment in the workshop that really doesn't drop the ball. This is what we understand as the foundation of "first tier".

Two, the two AIs aren't a gimmick — they save time for the workshop and procurement

Many MES also pin on an "AI", but click in and it's still old reports reskinned. We made two smart assistants that actually work separately:
The first is a "voice query assistant" for the workshop.
The shift lead doesn't flip menus, just asks: "Which looms had the longest downtime today" "Which workshop had the most warp breaks this month" — the system pulls real production data for you, can talk like a human, and can export tables. The voice recognition has hot-word bias for industry jargon like POY, yarn count, process step — you don't have to accommodate it.
The second is the "天机 Price Forecast" for procurement and the boss.
Cotton yarn, POY, FDY, DTY, nylon, spandex — these raw materials that shuttleless looms really use — we price them from multiple public real sources, cross-check, then give the price trend. Here's one iron rule: **if the data is collected for real we report the real price; varieties still under construction are honestly marked "under construction"; we never use mock numbers to fool people, and never decide for the customer "should you stock up"** — only reference, never investment decisions for you.
We'd rather say outright "we haven't done this one yet" than trick you with pretty numbers.

Three, full chain, from cotton all the way to apparel

Textile MES fears most "only managing one segment". The chain we want is complete:
Cotton spinning → Weaving → Dyeing → Apparel → Home textiles
Raw materials from yarn to greige, to dyeing-finishing, to apparel, to home textile products — every batch is traceable. Whether it's chemical fibers on water-jet looms, cotton yarn on air-jet looms, or raw materials used on large circular knitting machines — we're all on this chain.
Multi-factory, multi-language is also done — one system supports eleven languages, follows the factory wherever it goes and works.

Four, in closing

We don't dare claim we're "already number one in the world". But we dare say three things:
  1. The hardest pain point, we solved for real with hardware patents;
  2. Every number running in the system is real — if we can't do it, we say so;
  3. From cotton spinning to apparel, what we endorse is full-chain, real workshops — not demo animations.
There's no shortcut, no fast track in this trade. We're willing to grind workshop by workshop.
If you're also looking for a textile MES that "actually works after you buy it", welcome to chat — the demo site is always open, judge it yourself.