Structured Workflow Engine MCP Server

#!/usr/bin/env tsx import fs from 'fs'; import path from 'path'; import yaml from 'js-yaml'; import OpenAI from 'openai'; import { config } from 'dotenv'; // Load environment variables config(); interface WorkflowEmbedding { id: string; title: string; description: string; category: string; tags: string[]; complexity: string; searchableText: string; embedding: number[]; } const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY }); const WORKFLOWS_DIR = './public/playbook/workflows'; const OUTPUT_FILE = './public/workflow-embeddings.json'; async function buildEmbeddings() { console.log('🚀 Building workflow embeddings...'); if (!process.env.OPENAI_API_KEY) { console.error('❌ OPENAI_API_KEY is required'); process.exit(1); } try { // Ensure output directory exists const outputDir = path.dirname(OUTPUT_FILE); if (!fs.existsSync(outputDir)) { fs.mkdirSync(outputDir, { recursive: true }); } // Read all YAML workflow files const workflowFiles = fs.readdirSync(WORKFLOWS_DIR) .filter(file => file.endsWith('.yml') || file.endsWith('.yaml')); if (workflowFiles.length === 0) { console.error('❌ No workflow files found in', WORKFLOWS_DIR); process.exit(1); } console.log(`📄 Found ${workflowFiles.length} workflow files`); const embeddings: WorkflowEmbedding[] = []; for (const file of workflowFiles) { const filePath = path.join(WORKFLOWS_DIR, file); const workflowId = path.basename(file, path.extname(file)); console.log(`⚡ Processing: ${workflowId}`); try { // Read and parse YAML const yamlContent = fs.readFileSync(filePath, 'utf-8'); const workflow = yaml.load(yamlContent) as any; // Extract workflow metadata const title = workflow.name || workflowId; const description = workflow.description || ''; const category = workflow.category || 'general'; const tags = workflow.tags || []; const complexity = workflow.metadata?.complexity || 'medium'; // Build searchable text from all relevant fields const searchableText = [ title, description, category, ...tags, complexity, // Include phases and steps for better searchability ...(workflow.phases || []).map((phase: any) => { const phaseText = [ phase.phase || phase.name || '', phase.description || '', ...(phase.steps || []).map((step: any) => { return [ step.id || '', step.miniPrompt || '', step.description || '' ].join(' '); }) ]; return phaseText.join(' '); }) ].join(' ').toLowerCase(); // Generate embedding console.log(`🧠 Generating embedding for: ${title}`); const embeddingResponse = await openai.embeddings.create({ model: 'text-embedding-3-small', input: searchableText, encoding_format: 'float' }); const embedding: WorkflowEmbedding = { id: workflowId, title, description, category, tags, complexity, searchableText, embedding: embeddingResponse.data[0].embedding }; embeddings.push(embedding); console.log(`✅ Generated embedding for: ${title} (${embeddingResponse.data[0].embedding.length} dimensions)`); // Add small delay to respect rate limits await new Promise(resolve => setTimeout(resolve, 100)); } catch (error) { console.error(`❌ Error processing ${file}:`, error); } } // Save embeddings to file console.log(`💾 Saving ${embeddings.length} embeddings to ${OUTPUT_FILE}`); fs.writeFileSync(OUTPUT_FILE, JSON.stringify(embeddings, null, 2)); console.log('✅ Workflow embeddings generated successfully!'); console.log(`📊 Summary:`); console.log(` - ${embeddings.length} workflows processed`); console.log(` - Categories: ${[...new Set(embeddings.map(e => e.category))].join(', ')}`); console.log(` - File size: ${Math.round(fs.statSync(OUTPUT_FILE).size / 1024)} KB`); } catch (error) { console.error('❌ Error building embeddings:', error); process.exit(1); } } // Run the script buildEmbeddings();