Agent1 Commerce MCP / by Aumatron
Agent1 gives AI agents a secure, structured way to understand and interact with online stores. Search products, compare options, check availability, build a basket and hand the customer securely to checkout through one predictable MCP interface.
Most AI shopping integrations depend on page scraping, browser automation or platform-specific code. An agent must discover menus, forms, selectors and separate tools before it can complete a simple request. Agent1 replaces that fragile layer with a stable commerce interface.
A theme update, renamed field or changed selector can break the entire journey. Each AI application has to rebuild the same integration logic.
The AI sees the same tool names and response structure. Agent1 handles the platform-specific catalogue, stock, basket and checkout connections behind the scenes.
The shop keeps its existing catalogue, inventory controls and checkout. Agent1 adds the standardized connection that compatible AI agents can understand.
The business connects its ecommerce system and chooses which Agent1 capabilities to enable.
Aumatron or another compatible MCP client connects through the Agent1 service.
The AI searches, compares and checks live product information using structured tools.
With approval, the AI prepares a basket and sends the customer to the shop's checkout.
An AI should not need dozens of overlapping tools. Agent1 is designed around a compact set of clear operations with structured inputs and outputs.
CORE SCHEMA / IN DEVELOPMENTsearch_productsREADSearch by natural-language requirements, category, price and product attributes.
get_productREADReturn current pricing, variations, images, descriptions and structured attributes.
check_stockREADCheck whether the selected product and variation are currently available.
compare_productsREADPresent normalized product information for useful, accurate comparisons.
create_cartWRITECreate a store basket without committing to an order or taking payment.
add_to_cartWRITEAdd a selected product, quantity and variation to the active basket.
get_shipping_optionsREADReturn delivery methods and estimates when supplied by the connected shop.
create_checkout_linkHAND-OFFSend the customer to the shop's existing checkout for final review and payment.
Give AI applications a controlled way to search your live catalogue, answer product questions and prepare customer baskets without rebuilding your website for each AI platform.
For online businesses →Connect to a consistent commerce interface across supported stores. Spend less time maintaining browser selectors and platform-specific logic.
For AI developers →AI-assisted shopping should not bypass the store's existing security, payment or customer-confirmation processes.
Read and write operations are separated so each shop can control what an agent may access or change.
Basket changes and checkout hand-off are designed around clear customer approval.
Final review, address details and payment stay inside the shop's existing checkout flow.
Security and permission details shown here describe the intended product design and will be finalized before public availability.
Agent1 / Commerce MCP
Agent1 is being developed as the MCP layer between AI applications and ecommerce stores. The first release will focus on a reliable core toolset, clear permissions and secure checkout hand-off.