Work from the source.
Control logic, I/O configuration, and task definitions live in ordinary files. An agent can explore the same system your engineers do.
Put the machine’s behavior, history, and configuration where your team and your agents can work with them. Build capability that carries forward with every change.
Follow an illustrative change to the real motor-latch example: increase its start debounce from 200 to 500 milliseconds.
Ask your agent to trace the motor-start behavior. It can read the program, inspect the I/O mapping, and follow Git history—all in the same project.
You define the behavior you want.
Your agent gathers the context.
sealin/
├── plc.toml # identity and limits
├── programs/
│ └── Main.st # debounce and motor latch
├── io/
│ └── rio.toml # start, stop, motor mapping
└── tasks.toml # program scheduling
# Follow the history behind the behavior
git log -p -- programs/Main.st“Increase the start debounce to 500 milliseconds.” The agent proposes a small edit to the Structured Text. The exact behavior change is visible in a normal Git diff.
You review the proposed change.
Your agent edits the source.
PROGRAM Main
...
- debounce(IN := start, PT := T#200ms);
+ debounce(IN := start, PT := T#500ms);
seal := (debounce.Q OR seal) AND NOT stop;
motor := seal;
END_PROGRAM
# Review every changed line
git diff -- programs/Main.stValidate the project, compile an image, and exercise the behavior in simulation. The same command-line tooling is available to your engineers and your agents.
You decide what good looks like.
Your agent runs checks and helps investigate.
# Validate project configuration
thermite check ./sealin
# Compile the control program
thermite build ./sealin -o sealin.tplc
# Inspect the resolved configuration
thermite inspect ./sealin
# Then test timing and stop behavior
# in your configured simulation environment.After you approve and deploy the change, observe live values and inspect diagnostics. Commit the change and its rationale so the next person—or agent—can pick up the thread.
You control the deployment decision.
Your agent helps interpret the evidence.
# Stream values as machine-readable samples
thermite monitor %IX0.0 %QX0.0 --json
# Inspect runtime and device health
thermite status
thermite diag
# Open the browser diagnostics
thermite web
# Keep the change and its rationale in Git.Bring your own coding agent. ThermitePLC provides the readable project and command-line tools.
A source repository can preserve more than the latest program. Commit messages explain intent. Diffs show decisions. Simulation code captures expected behavior.
An agent can use that context to help a new engineer understand what came before—and make the next change with a clearer view.
Logic, I/O mapping, and configuration.
A reviewed change. Its reason, recorded.
The same source. The context to keep going.
Control logic, I/O configuration, and task definitions live in ordinary files. An agent can explore the same system your engineers do.
Use Git diffs to understand what changed. Keep the reasoning beside the implementation, where the next engineer can find it.
The CLI gives both people and agents access to compilation, inspection, monitoring, and diagnostics. Results feed the next decision.
The project is the shared context. Your process decides how changes move from an idea to a machine.
Explore a PLC projectYour control logic. Your project. A PLC license that fits.
Take the next step with ThermitePLC.