In the rapidly evolving landscape of agentic AI, many teams building shopping assistants or merchant tools find themselves recreating the same foundational scaffolding: an agent loop, a tool layer for catalog integration, approval gates, and evaluation suites. Recognizing this redundancy, Anthropic has now released that scaffolding as an open-source blueprint. In September 2026, the company published anthropics/commerce-agents on GitHub, providing a comprehensive reference implementation that includes both a shopping agent and a merchant agent, along with four runnable verticals: retail, travel, telecom, and entertainment. The release is accompanied by a product announcement and an engineering deep-dive titled A guide to the anatomy of effective commerce agents.
Is it deployable? Absolutely. The repository is licensed under Apache 2.0, runs locally on Python 3.11+ and Node 22 with an ANTHROPICAPIKEY, and the runtimes accept any anthropic client. This means the same codebase can be deployed seamlessly across the Claude API, Amazon Bedrock, Microsoft Foundry, or Google Cloud Vertex AI.
The Two Agents
The shopping agent is designed to operate within a merchant's own application. It searches the catalog, handles multi-item requests, compares options, builds the cart, and answers order and returns inquiries in the same conversational flow. Its five core skills are search-discovery, purchase-research, planning-goals, customer-care, and memory-personalization. To deploy, a merchant implements a StorefrontBackend interface that connects to their catalog, cart, order, and policy systems.
The merchant agent, on the other hand, supports store staff by answering questions about sales performance, providing inventory alerts, offering pricing and promotion recommendations, and assisting with customer-related tasks—essentially serving as an intelligent backend assistant for retail operations.
via MarkTechPost