What's new in Fluid Network Studio, September 2026

Two additions this month change how a network gets onto the canvas and how it looks once it is there. Both are free to try.

Build a network with your AI assistant

Describe the system in plain English to the AI assistant you already use. Open File then Build with AI in the Studio, press Copy prompt, and paste it into your assistant along with what you need: the lengths, the elevations, the flow you are designing for. Your assistant writes the network as a file, you paste its whole reply back, and the Studio digs the file out of the prose around it and checks it before anything reaches your canvas.

The checking is the point. Every dimensional value has to carry its unit, so a diameter meant in millimetres can never land in a metres field. Every pressure has to say gauge or absolute. Pipe materials, fittings and fluids are matched against the Studio's own libraries, and an unknown name is an error that lists the nearest valid ones. Unknown ids, duplicate ids, a fitting without a pipe on both sides, a part with no boundary, a pump in a gas network: all caught before anything loads.

You see a review first, listing what is in the network, what your assistant assumed and anything it got wrong. Every message is written so you can paste it straight back, and there is a button that copies the whole list. Changing a network works the same way: ask for the change, paste the reply, read what was added, changed and removed, then apply it in one undo step.

Your network never goes to an AI company through us. You choose the assistant and what you share with it, and the guide we publish for assistants is public, versioned and machine-readable, so any assistant can read it.

AI output can be wrong. The importer checks structure, units and connections, not engineering judgement, so check every value before you rely on the results.

Read how it works, with a worked example.

Draw pipes to scale

A new To scale toggle in the Studio toolbar draws every pipe in proportion to its entered length. A 1 m pipe is half the length of a 2 m pipe on screen, a scale bar in the corner tracks your zoom and reads in your own length unit, and pipes you draw while it is on follow the same rule. It is the fastest visual check of entered lengths there is: a 1000 m pipe typed where you meant 100 m stands out immediately.

Your schematic layout is kept exactly as you left it. Switch back and every symbol is where you put it, down to the pixel. Lengths stay entered data and are never read back from the drawing, so turning To scale on can never change a solve input or mark your results stale.

Loops whose lengths cannot close with straight lines, and pipes too short to show at the current scale, are drawn as close as they can be and flagged, with a chip that highlights them on the canvas.

Also this month

• Solving now follows the same rules however a network reached the canvas, so a file you open or an EPANET model you import behaves exactly like one you drew. The Solve button and the command palette now agree, which they did not before.
• Signed-in users can find these notes any time from the account menu, under View updates.

The full record of changes, month by month, is on the changelog.