workflow-interactive-dev

Interactive nodes in FastGPT let you pause workflow execution to collect user input or decisions. This Skill guides you through defining node types (userSelect, formInput, childrenInteractive, loopInteractive, paymentPause), implementing backend dispatch logic, and building frontend components—from type definitions to full integration.

Interactive nodes require type definitions, templates, nodeOutputs, memoryEdges, skipNodeQueue flags, and isEntry handling. The Skill details how these components work together in the workflow service to manage node dispatch, history rewriting, and runtime behavior across backend and frontend layers.

Yes. Use formInput nodes to gather multi-field user data or paymentPause nodes for approval checkpoints. This Skill provides the full development path: defining the node type, creating templates, implementing backend logic, and building the React frontend component for seamless workflow integration.

You'll need TypeScript and React knowledge, understanding of FastGPT's workflow service architecture, and familiarity with node dispatch and history handling. The Skill covers isEntry flag management, testing checklists, and functional requirements like memoryEdges to ensure proper behavior.

Interactive nodes can modify workflow history when user decisions branch execution. This Skill explains the orchestration between node dispatch logic and history rewriting in the workflow service, ensuring state consistency and proper playback when workflows pause and resume.

userSelect lets users choose predefined options, formInput collects structured data, childrenInteractive and loopInteractive handle nested or repeated interactions, and paymentPause pauses for transactions. The Skill details each type's architecture, runtime behavior, and implementation requirements.