Introducing Oyster’s MCP server.
Give the assistant a defined job
Give your AI assistant one clear task. For example, ask it to explain why a product matches a customer’s skin report.
Where MCP fits
MCP lets AI assistants connect to tools and data. Oyster offers it alongside the API and widget. The widget provides the scan screen. The API returns skin data and product matches. MCP is the connection for AI assistants.
Begin with one conversation
Start with one task, such as explaining a product match with the customer’s consent. Use the scan results and product catalogue. Never invent missing scores.
Agree the connection details
Ask Oyster for your connection address, sign-in method and available tools. Check which tools can read data and which can change it. Access can vary by account.
Make the limits visible
Plan what happens when a result is missing, a product is unavailable or the connection fails. Give the customer a way to reach a person.
Check the real workflow
Test the connection with your team before launch. Check access permissions, the data the assistant sees and how failed requests are handled.
Plan the integration
Start with the developer overview and agentic commerce in beauty. To establish access and a first use case, contact the Oyster team.
See how Oyster fits your business.
Bring your storefront, your questions and the workflow you want to improve.