Packages

@loop-engine/adapter-commerce-gateway

Install

1npm install @loop-engine/adapter-commerce-gateway

Install one LLM SDK based on provider choice:

1npm install @anthropic-ai/sdk
2# or
3npm install openai

Prerequisites

• COMMERCE_GATEWAY_URL
• COMMERCE_GATEWAY_API_KEY

CommerceGatewayClient

1const client = new CommerceGatewayClient({
2  baseUrl: process.env.COMMERCE_GATEWAY_URL!,
3  apiKey: process.env.COMMERCE_GATEWAY_API_KEY!
4})

Methods:

• getInventory(sku)
• getInventoryBatch(skus)
• getDemandForecast(sku, horizon?)
• getSuppliers(sku)
• getCurrentPrice(sku)
• getPriceHistory(sku, days)
• createPurchaseOrder(order) (write operation; post-approval only)
• recordLoopOutcome(outcome) (optional endpoint support by deployment)

buildProcurementActor

1const actor = buildProcurementActor({
2  gatewayClient: client,
3  llmProvider: 'openai',
4  apiKey: process.env.OPENAI_API_KEY!,
5  confidenceThreshold: 0.8
6})

Returns async (context) => Evidence, where context.instance.data.sku drives Gateway queries and the actor returns evidence for a guarded transition.

Evidence shape

1type ProcurementEvidence = {
2  recommendedSku: string
3  recommendedQty: number
4  estimatedCost: number
5  supplierId: string
6  confidence: number
7  rationale: string
8  gatewayRequestIds: string[]
9  modelUsed: string
10  timestamp: string
11  _confidence: number
12}

Read vs write boundary

AI actor functions in this package call read endpoints only. Order creation stays in automation flows after a human approval transition passes guard checks.

buildPricingActor

Related example

See LLM Commerce Gateway Integration for the end-to-end procurement flow.