Platform Capabilities v2

The Runtime provides synchronized visual and YAML authoring, compiled execution, Cluster placement, production triggers, Flowgear Apps, and agent-assisted delivery.

Design and connectivity

Build a Workflow as ordered or nested Steps. The canvas and YAML definition edit the same source, while validation checks Node contracts, types, mappings, Expressions, triggers, and Connections before compilation.

Connections hold endpoint and authentication values separately for each Environment. The runtime validates that the selected Connection matches the Node, version, and Cluster placement required by the Step.

Execution and placement

The runtime compiles a Workflow and places the whole Workflow on one cloud or Local Cluster. Streams and asynchronous collections can be consumed lazily when the participating Steps preserve deferred processing.

See Runtime and Lazy Evaluation and Streaming.

Triggers and interfaces

Run a Workflow in Debug or enable supported trigger Steps for schedules, listeners, HTTP endpoints, and MCP tools. A Workflow can also be invoked as a Sub-Workflow or used behind a Flowgear App.

Use Different Ways to Run a Workflow for the exact trigger and activation requirements.

Operations and governance

Workflow Logs show Step status, timing, counters, Property previews, Returns, and exceptions. Property-level redaction can prevent selected values from being captured.

Revision Management retains source history, Release Management promotes revisions through Environments, and Environment-specific Workflow status controls whether published triggers can accept new starts.

Apps and agent-assisted delivery

Flowgear Apps provide lightweight web interfaces backed by authorized Workflows. The AI Workflow Assistant works inside the designer, while the Builder MCP Server exposes the Workflow-building tool surface to compatible external coding agents.