# Runtime (/developers/pinets/architecture/runtime)



## PineTS Runtime Engine [#pinets-runtime-engine]

The runtime environment executes the transpiled code. It simulates the Pine Script execution model where the script runs once for every bar in the dataset.

## Key Components [#key-components]

* **PineTS Class**: The entry point. Manages the data provider, pagination, and execution loop.
* **Context Class**: The "Global State" passed to the script. Holds all variables, series data, and namespace instances.
* **Series Class**: The fundamental data structure wrapping arrays.

## Execution Model [#execution-model]

PineTS follows a "Push-Based" execution model:

1. **Load Data**: Fetch OHLCV data from a Provider.
2. **Initialize Context**: Create the `$` object.
3. **Run Loop**:
   * For each historical bar:
     * Update `context.idx`.
     * Push new OHLCV data into `context.data` arrays.
     * Execute the user script.
     * Grow all user variables (push current value to history).
   * Collect results/plots.

This ensures that `close[1]` always refers to the data from the previous iteration.