Lehrling now has made the canvas for engineering workflows open source — currently in beta. Join the waitlist.

Automate workflows. Keep the expertiseMANUFACTURING INTELLIGENCE CANVAS

Turn your shopfloor operations into unified visual workflows. Lehrling handles the repetitive checks—your experts keep total control over every call.

Quoting

RFQ packet

Documents

CAD analyzer

Cost model

CAM

CAM software

NC simulation

Requirements

Machining operations

Setup

Tools

Cycle

machine

Inspection card

Drawing

Documents

Open questions

Industry focus

HMLV job shopsMold & diePrecision millAerospaceMedical

Turn an RFQ packet into a first-pass quote

RFQ packet

PDF · drawing

● Ready

Analyze Documents

Role + output

● Live

CAD Analyzer

STEP / STL

● Live

Cost Model

Hours × rates

● Quote

Plug into your existing software with 20+ integrations

CAD, CAM, ERP, MES, and the machine already on the floor.

Fusion 360
Onshape
SolidWorks
CATIA
Siemens NX
Inventor
Mastercam
hyperMILL
PowerMill
TopSolid
Creo
Autodesk
SAP
FANUC
Mazak
Heidenhain
Epicor
PTC
ANSYS
Rockwell
ABB
SketchUp
Dassault Systèmes

Build AI agentswhere it actually matters

Use AI agents to solve problems that they are actually good at.

Build complex AI without getting boxed in

Multi-step quote graphs, document roles, CAD in the browser. Swap how a node thinks without rebuilding the floor process around it. Author a component when the palette is not enough.

Runs where the work already happens

Kai watches the desk and the floor. Knowledge stays with the part, not in a slide deck. Use the hosted demo today; shop-owned deployment is the point of a pilot.

Let people and logic guide AI decisions

Give Analyze Documents a role and an expected output. Keep estimators on the cost model. Inspect every execution: the file, the prompt shape, and what Kai returned.

Move fast. Break nothing.

Short feedback loops that keep quoting and programming in flow.

  • Re-run a single node — not the whole quote graph
  • Replace a part or document without resetting the canvas
  • Default template already wires CAD Analyzer to Cost Model
  • Live STEP/STL analysis in the browser, nothing uploaded for geometry

Case studies

RFQ → first-pass quote

How estimators close RFQs without waiting on a senior

“The canvas is the unlock. Chat is useful, but PathCraft is what lets you put CAD, documents, and cost on one path you can actually explain.”

Estimating lead · precision job shop

Session → shop memory

How CAM knowledge stops walking out with one programmer

“Kai sitting beside the CAM session is how the next person gets the why — not just the toolpath file.”

Programming manager · mold & die

Enterprise-ready

Reliable. Scalable. Secure.

Deploy the way a shop actually buys software. Watch agents without losing control of the quote.

Security and control

Invite-gated demo, signed-in sessions, shop data stays in your conversation and files — not scraped into a public model card.

Observability

Canvas status on every live node. Document analysis you can re-run. Chat history you can read.

Builder experience

Drag-and-drop graph, custom components, CAD and LLM nodes in the same editor.

AI governance

Role prompts, expected-output contracts, human markup on cost. Kai does not silently ship a number.

“I don't want another black-box copilot. I want to see the graph that priced the part.”

Owner-operator

“Dropping the STEP on the canvas and watching features land is the first time AI felt like a junior on my bench.”

CAM programmer

“The role field on Analyze Documents is the whole product. I tell it to be a quality engineer. It behaves like one.”

Quality lead

“We don't have time to train. If it can't learn while we quote, it doesn't get on the floor.”

Shop supervisor

“Custom nodes mean I'm not waiting on a vendor roadmap for fixture planning.”

Process engineer

“Cost model next to CAD is how quoting should have worked ten years ago.”

Estimator

Simple enough to see. Powerful enough to ship.

Join the shops building AI automation they can actually explain — on the canvas, not in a prompt graveyard.

nitin@lehrlingai.com