Operate inspection equipment in everyday language and keep veterans' instructions and judgments as a knowledge base. Runs on in-house GPUs. New in 2026.
Operate by speaking, and the know-how becomes a company asset
Give instructions in everyday language and the AI sets up the equipment, keeping veterans' instructions and judgments in a knowledge base and inspection packages for the next operator. People change; the know-how stays.
The three safety gates, in-house GPUs and zero external communication are as shown in the company product video.
Videos show camera exposure, lighting and focus adjusted in natural language.
Everyday instructions: “Make the lighting a little brighter and inspect it again.”
Three safety gates: Deterministic decomposition · human approval · read-back confirmation
Process data security: Runs on in-house GPUs with zero external communication
Know-how as an asset: Veterans' instructions and judgments → knowledge base and inspection packages → the next operator
Natural-language commands: Give orders in plain words, such as “raise light channel 2 by 20% and run one inspection”, and preview them before they run.
One-click commands: Station status summary, a single inspection run, lighting adjustments, undoing the last adjustment and emergency stop, each with one click.
Approval by risk level: Questions, inspections and small lighting changes run at once; high-risk commands such as camera or recipe changes wait for a person’s approval.
Safety limits (guards): Adjustments stay within the minimum and maximum set for each device, and current values can be checked at any time.
Inspection station link: See the IntScan Pro station’s connection status, recipe, lighting and latest result on one screen.
Command history: Every command is logged in order with whether it ran and its result, so you can trace which command changed what.