Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
import type { CompletionParams, Model } from '@/types/app'
import { createLocalStorageState } from 'foxact/create-local-storage-state'
import { ModelModeType } from '@/types/app'
export const EMPTY_WORKFLOW_GENERATOR_MODEL: Model = {
name: '',
provider: '',
mode: ModelModeType.chat,
completion_params: {} as CompletionParams,
}
const [useWorkflowGeneratorModel, _useWorkflowGeneratorModelValue, _useSetWorkflowGeneratorModel] =
createLocalStorageState<Model>('workflow-gen-model', EMPTY_WORKFLOW_GENERATOR_MODEL)
// Last instruction the user generated from, persisted across opens so reopening
// the generator resumes where they left off instead of a blank box (the palette's
// inline-captured instruction still takes precedence over this).
const [
useWorkflowGeneratorLastInstruction,
_useWorkflowGeneratorLastInstructionValue,
_useSetWorkflowGeneratorLastInstruction,
] = createLocalStorageState<string>('workflow-gen-last-instruction', '')
export { useWorkflowGeneratorLastInstruction, useWorkflowGeneratorModel }