Let’s connect
OPEN SOURCE. OPEN POSSIBILITIES.
Your system has a lot of moving parts.
Get them working
together.
Plugboard connects your Python models into powerful simulations and workflows—so you can focus on how your system works, and what could make it better.
01 / THE PROBLEM
Real systems don’t run in straight lines.
Machines fail. Material loops back. Decisions change the next step. Connecting all of that with custom glue code can become harder than modelling the process itself.
Plugboard takes care of communication between components and their execution. You define the behaviour and the connections. Data moves through the system as events happen.
Meet the building blocksTurn your Python logic into reusable components. Mix physics, machine learning and live data in the same process.
Define how components exchange data, including branches and feedback loops. Plugboard handles the flow.
Run locally, change parameters, compare scenarios, then scale out with Ray when your model needs more compute.
02 / POSSIBILITIES IN PRACTICE
Different domains. The same need to understand how the parts work
together.
Explore examples from the Plugboard docs.
Build a digital twin of a production line. See how machine failures, buffers and maintenance schedules affect the whole operation before changing the real thing.
Connect crushers, screens and conveyors into a dynamic process model. Follow material as it circulates, and explore how operating settings change the final product.
“What happens to product quality when we change the crusher settings?”
Material in. Insight out.Bring physical processes and controllers into one model. Explore changing demand, coupled tanks and treatment processes with feedback between components.
“How does our system respond when demand suddenly changes?”
Physics meets feedback.Model the interactions that turn small delays into network-wide effects. Test changes to traffic signals and bus operations in repeatable scenarios.
“Why do our buses arrive together—and what could keep them apart?”
Local decisions. Network effects.Compose market-data sources, signal calculations and portfolio logic as separate components. Replay data and compare strategy behaviour in a reproducible research workflow.
“How does this strategy behave when we change the signal?”
Repeat. Compare. Understand.Make LLMs part of a wider working system. Connect model providers to data sources, transforms and downstream components, and swap providers as your requirements change.
“What if an AI model could work alongside our existing simulation?”
Intelligence, connected.Your domain isn’t on the list? If it has interacting components, let’s talk. Tell us what you’re modelling
✳ 03 / BUILT FOR THE WAY YOU BUILD NOW
Describe the system you have in mind.
Give your AI assistant
the tools to build it.
Run plugboard ai init to install project context and
agent skills. Your coding assistant gets specific guidance to
create components, connect models, generate diagrams and run
scenarios.
You stay in control: inspect the Python, review the assumptions and validate the results.
Explore AI-assisted development$ python -m pip install plugboard
$ plugboard ai init
“Model a production line with three machines, random processing times and occasional failures. Compare throughput with different buffer sizes.”
Editable Python. Reusable YAML. Scenarios you can run.
Build with AI. Build AI in. Plugboard also includes LLM components, so AI can be part of the model itself.
04 / YOUR NEXT CONNECTION
Open source. Python-native. Ready for your first component.
$ python -m pip install plugboard
Python 3.12+ · Apache 2.0 licence · Local to cloud with Ray
LET’S CONNECT
A tricky process, an idea for a digital twin, or a question about getting started. We’d love to hear about it.
hello@plugboard.dev