Connect the physics. Decide what converges.
Flow is our graphical workflow environment. Place solvers on a canvas, wire them in series or in self-consistent loops, and let the built-in AI agent do the setup. All on your own machine.
Place. Connect. Run.
Place
Drop Transport, Optical (CMT, FEM, FDTD, ray tracing) and Thermal solvers on the canvas. Add external tools as nodes beside them.
Connect
Draw arrows between nodes. A single arrow passes results one way. A double arrow iterates two solvers until they agree.
Run
Run the whole graph, follow what is happening, and inspect the results of every node.
One arrow or two. You choose.
Not every problem needs full self-consistency. Flow lets you spend computation only where the physics feeds back.
Single arrow: one-way
B receives the result of A, once. Fast, and ideal when the coupling is weak or runs in one direction only.
Double arrow: self-consistent
A and B iterate until they agree. Use it where the physics closes a loop, such as self-heating that shifts the optical response, which in turn changes the heat.
Same solvers, different questions.
Fast design pass
Modes first, then the temperature they produce. One pass, in seconds.
Full chain in series
Add carrier transport at the end of the chain without paying for full iteration.
Everything self-consistent
All three physics converge together, for the most predictive operating point.
Absorption to current
Optical absorption feeds the carrier transport model.
Self-heating
Current heats the device and temperature changes the current, until both settle.
Your own graph
Build the chain your device needs, including tools from outside Po.Li.Te.
Describe the study. The agent builds it.
Flow includes an AI agent that works inside your project, not outside it.
- Suggests how to set up the simulations for your device and your question.
- Writes the connections, the simulation flow and every input file it needs.
- Runs the simulations and keeps track of what is happening along the way.
It runs on a compact language model on a standard consumer graphics card. No data is sent to an external service.
Bring your own tools
Flow can integrate external tools, so a Po.Li.Te solver can sit next to the codes your team already trusts. Tell us which tools you need to connect and we will plan it with you.
Local by default
Flow, the solvers and the AI agent run on your hardware, even offline. If you prefer not to install anything, we offer a cloud solution hosted in the European Union.
Questions about Flow
What is Flow?
Flow is the graphical workflow environment of Po.Li.Te. You place solvers (transport, optical, thermal) and external tools on a canvas, connect them, and run the resulting simulation flow.
What is the difference between a single and a double arrow?
A single arrow passes the result of one solver to the next in one direction. A double arrow makes two solvers iterate until their results are self-consistent. You can mix both in the same flow.
Does the AI agent send my data to the cloud?
No. The agent runs on a compact language model on your own machine, and a standard consumer graphics card is enough. Your structures and results do not leave your hardware.
Can Flow connect external tools?
Yes. Flow can integrate external tools as nodes, so Po.Li.Te solvers can work together with the codes your team already uses. Contact us to discuss the tools you need.
Can I deploy Flow in the cloud?
Yes. Flow runs locally by default. If you prefer, we offer a cloud solution hosted in the European Union.
Build your first flow with us.
Book a free demo and a 14-day trial. We set up a flow for your device together.