Modeler AI

PracSim Modeler AI puts a multi-provider AI chat panel inside FlexSim. Hook it up to your Anthropic, OpenAI, or other provider API key, and the assistant can build, run, and discuss your simulation directly, with the active model loaded as context. It is open source and BYOK (bring your own key): you choose the model, you pay the provider, and your conversations never touch PracSim's servers.

What you get

• Build, run, and discuss your model in chat without leaving FlexSim
• Work remotely: connect to your model and build it from another device, away from the computer it runs on
• Model context loaded automatically, so the assistant works against the active simulation
• BYOK: works with Anthropic Claude, OpenAI GPT, and other providers
• Your API key, your provider, your data. No PracSim relay.
• Open source. Free to use, free to extend or fork.

What it looks like

Build it in chat

Ask for what you want and Modeler AI works through it step by step. Here it is planning, then building, a small production line (a Source feeding three Processors into a Sink) with process times, then a "Line Performance" dashboard with throughput, WIP content, utilization, and staytime charts. The assistant explains the plan, creates the objects, wires the flow, configures the dashboard, runs the model, and reports back.

Your model, with the dashboard it just built

The same conversation, after the run completes. The 3D view shows the production line the assistant built. The "Line Performance" dashboard shows the throughput, WIP content vs time, average staytime, and utilization by hour from a 10,000-second run. The chat panel summarises what was built and offers refinements.