Nutrient Document Engine MCP Server

/** * Procurement Workflow * * Orchestrates the document processing pipeline for procurement automation. * The workflow processes documents through three sequential stages: * 1. Upload - Uploads documents to the Document Engine * 2. Classify - Uses AI to classify documents as PO, Invoice, or Payment * 3. Match - Performs intelligent 3-way matching between related documents */ import { END, START, StateGraph } from '@langchain/langgraph'; import { ConfigurationSchema, WorkflowStateAnnotation } from './State'; import { classifyDocumentsNode } from './nodes/ClassificationNode'; import { matchDocumentsNode } from './nodes/MatchingNode'; import { uploadNode } from './nodes/UploadNode'; export const procurementWorkflow = new StateGraph(WorkflowStateAnnotation, ConfigurationSchema) .addNode('upload', uploadNode) .addNode('classify', classifyDocumentsNode) .addNode('match', matchDocumentsNode) .addEdge(START, 'upload') .addEdge('upload', 'classify') .addEdge('classify', 'match') .addEdge('match', END) .compile();