LADX
Start free
← Products

Knowledge

Available now

Your manuals, answering questions.

Drop in the drive manual, the machine spec, the site standard. Ask questions against them and get answers with the page they came from.

Open Knowledge
Every answer carries the page behind it“What is the maximum torque limit on the drive?”VLT-FC302 p.184page 184the other 400
An answer is only useful if you can check it. Every one arrives with the document and page behind it.

In short

Can I ask questions of my own PLC and drive manuals?

LADX Knowledge indexes documents you upload, drive manuals, machine specifications, site standards, and answers questions using only passages from them, with the page each answer came from. Anything not in the documents comes back as not in the documents rather than as a plausible invention.

The problem

The answer is almost always in a PDF somebody has. Finding which PDF, and which of its four hundred pages, is the actual work.

What it does

  • Answers with citations

    Every claim points at the document and the page it came from, because an unsourced answer about a torque limit is worse than no answer at all.

  • Scoped per project

    One project's manuals never leak into another's answers, and nothing is used to train anything.

  • It works with the rest

    Chat can search it while writing logic, so a generated rung can respect the limits stated in the drive manual.

What it doesn't do

Listed because you would find out anyway, and finding out later is worse.

  • It retrieves and quotes. It does not verify that the manual is the current revision; that is still on you.
  • A scanned manual with no text layer is a picture. It needs a PDF with real text in it.

Questions

How is this different from asking ChatGPT about a manual?
It answers only from passages retrieved out of your documents, and shows which passages it used. A general model asked about a specific drive parameter will produce a confident number from nowhere in particular, and there is no way to tell from the answer which kind you got.
What happens if the answer is not in my documents?
It says so. That is the behaviour worth having: a retrieval system that pads a gap with general knowledge is more dangerous than one that refuses, because the refusal is visible and the padding is not.
What file types can I upload?
Text and PDF today. Extraction from PDFs with heavy tables loses some structure, which matters because tables are where parameter values live, and improving it is on the plan.