Reference · lookup

Look up any building block

Every winkComposer node grouped by category — parameters, inputs, outputs, one entry per node.

Nodes by Category

Clean and prepare raw data — smooth noise, filter frequencies, validate ranges, and remove spikes before they reach the pipeline.

Extract patterns from streaming data — track trends, measure correlation, detect state changes, and compute windowed statistics.

Spot anomalies and confirm them — detect mean shifts, check thresholds, validate persistence, and measure process capability.

Simple math between fields — running totals, differences, ratios, and flag inversion.

Control what runs — gate messages with conditions and coordinate nodes dynamically.

Get data out — broadcast alerts to MQTT and persist records to storage when conditions are met.

Looking for a specific node? See the Node Index.

Source adapters, emitters, storage, naming policies, and asset class binding — everything between flow() and .run().

Column schemas, three-tier limits, asset classes, enumerations, and the loader API — the complete metadata specification.

Dynamic options that adapt at runtime — built-in helpers, resolution behavior, and what happens when a tunable throws.