Claude Commerce Agents - Anthropic's open-source e-commerce Agents

What are Claude Commerce Agents

Claude Commerce Agents are Anthropic's open-source e-commerce Agents, including shopping and merchant Agents. The tools cover four industries: retail, travel, telecom, and ticketing, with built-in multi-layered safety guardrails, including staging all writes for approval and keeping the model away from payments. Claude Commerce Agents have been measured to increase shopping cart size by 35% and conversion rate by 60%. The tools support three deployment methods: Messages API, Agent SDK, and Managed Agents, and can integrate with platforms such as AWS Bedrock and Google Cloud.

Main features of Claude Commerce Agents

• Shopping Agent: Embedded in a merchant's app or website, it helps consumers search, compare prices, plan shopping lists, and build shopping carts using natural language.
• Merchant Agent: Provides operations staff with sales analysis, inventory alerts, pricing and promotion suggestions, and marketing campaign drafting capabilities.
• Multi-industry coverage: Provides four runnable vertical industry reference implementations: retail, travel, telecom, and ticketing.
• Safety guardrails: Prices and inventory are strictly anchored to real catalog data, all merchant-side writes are staged for manual approval, and the model does not touch the payment process.
• Three deployment methods: Supports Messages API, Agent SDK, and Managed Agents.
• Claude Code plugin: Automatically generates project code, business flows, and test suites through scaffolding commands, completing Agent setup within hours.
• Memory and personalization: Remembers user preferences across sessions and automatically extracts and updates user profiles to tailor recommendations.
• In-conversation UI: Product cards, comparison views, and shopping cart status are rendered directly in the conversation interface.

Follow WeChat and reply "open source" to joinAI Open Source Project Exchange Group

How to use Claude Commerce Agents

• Fork the repository: Clone anthropics/commerce-agents locally, and install the dependency environment for Python 3.11+ and Node 22.
• Configure keys: Copy .env.example as .env and fill in ANTHROPIC_API_KEY。
• Run examples: Execute python scripts/run_demo.py retail to launch interactive demos for four industries including retail.
• Integrate with backend: Implement the StorefrontBackend or MerchantBackend interface to connect the Agent to your catalog, shopping cart, inventory, and order systems.
• Claude Code scaffolding: After installing the plugin, run /scaffold-commerce-agent to automatically generate project code adapted to your tech stack.
• Choose deployment method: Customize the loop via Messages API, quickly start with Agent SDK, or deploy managed with Managed Agents.
• Toggle feature modules: Use enable_* to configure and disable unwanted capabilities (for example, if there is no shopping cart, disable the related tools and prompts).
• Custom Skills: Under skills/ , create a new directory with SKILL.md to extend business flows such as product discovery and customer service.
• Validation and testing: Run pytest and scripts/smoke_chat.py Ensure the Agent behaves as expected.

Core advantages of Claude Commerce Agents

• Ready to use, fast to deploy：Provides a complete code repository that can be forked directly, allowing engineering teams to integrate shopping or merchant Agents into their own systems and launch within days.
• Flexible deployment, multiple paths: Supports three deployment methods—Messages API, Agent SDK, and Managed Agents—and can seamlessly connect to mainstream cloud infrastructure such as AWS Bedrock and Google Cloud.
• Faster development, plugin acceleration: The Claude Code plugin shortens Agent setup from days to hours through scaffolding commands, and automatically generates test suites based on your own catalog and rules.
• Simple architecture, no redundancy: Adopts a "single Agent + Skills" architecture, avoiding the state handoff loss and extra latency caused by multi-Agent orchestration in strongly context-continuous scenarios such as shopping.
• Security built in, enforced in code: Fences, source gating, permission limits, and approval gates are all enforced at the code layer, ensuring security without relying on model self-discipline.
• Payment isolation, zero risk: The model is only responsible for building the shopping cart; the checkout URL and payment operations are handled entirely by the merchant system, and the model cannot see any payment information.
• Changes staged, human review: All merchant-side write operations are staged by default and take effect only after manual approval, preventing AI errors from directly affecting live business.

Project address of Claude Commerce Agents

• Project official website：https://claude.com/solutions/commerce
• GitHub repository：https://github.com/anthropics/commerce-agents

Comparison of Claude Commerce Agents with similar competitors

Application scenarios of Claude Commerce Agents

• Intelligent shopping guidance for e-commerce platforms: Consumers describe their needs in natural language, and the Agent automatically searches, compares prices, plans, and builds a shopping cart.
• Merchant Operations Decision Support: Operations staff inquire about sales performance and inventory status; the Agent proactively warns of slow-moving/stockout risks and recommends pricing and promotion plans based on historical data.
• Travel Itinerary Planning: The user raises multi-dimensional travel needs; in a single conversation, the Agent integrates price comparison, scheduling, and value analysis to generate a complete itinerary plan.
• Intelligent Telecom Plan Recommendations: Based on the user's account usage, budget, and number of lines, the Agent analyzes the plan matrix and discloses regulatory fees to recommend the optimal contract combination.
• Ticketing and Event Services: Handles limited-time bookings for concerts/sporting events, waitlist queuing, ticket transfers, and venue seat map navigation, while ensuring fee transparency and real-time inventory synchronization.