> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.mitu.ai/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.mitu.ai/_mcp/server. # Bring your product to AI agents Connect your existing API to Mitu and make selected capabilities available as tools that AI agents can discover and use. You do not need to build a separate backend or recreate the business logic that already powers your product. Your API continues to handle the same products, customer accounts and business logic it handles today. You choose which operations agents can use and how customers authenticate when they connect from an AI client. > **Agent Wallets are optional** > > You do not need to issue Agent Wallets to connect your product to AI agents. Agent Wallets are only needed when you want to give your customers wallets that their AI agents can use to make payments on their behalf. > > Learn more about [Agent Wallets](/platform/agent-wallets/overview). ## Before you begin You need a Mitu account, an existing API that Mitu can reach and at least one API operation that you want AI agents to use. If your tools need to identify the customer using them, you will also need to choose how customers authenticate. You can use customer keys or connect your existing OAuth sign in flow. #### Connect your existing API Start by connecting the API that already powers your product. In the dashboard, open your MCP configuration and provide the base URL that Mitu should use when calling your API. ```text https://api.example.com ``` You do not need to create separate endpoints for AI agents. The same API that serves your existing application can also handle requests initiated through MCP. When you connect your API, Mitu provides a signing secret that your backend uses to verify requests sent by Mitu. Keep this secret secure and use it only from your server. Learn more about [Request verification](/platform/security/request-verification). #### Choose the operations agents can use Choose which API operations you want to make available to AI agents. Each operation becomes an MCP tool that an agent can discover and call. For example, an existing commerce API might make these operations available. ```text GET /products GET /products/{id} POST /orders ``` Describe each operation clearly so an AI agent can understand what the tool does, when it should use the tool and what information it needs to provide. ```yaml name: create_order method: POST path: /orders description: Creates an order for the signed-in customer using the confirmed cart and delivery details when_to_use: Use after the customer has confirmed the items, quantities, delivery address and final price requires_auth: true ``` The description and schemas you provide help AI agents choose the correct tool and construct the request correctly. If you already maintain an OpenAPI specification, you can import it instead of configuring each operation individually. Learn more about [MCP endpoints](/platform/mcp/endpoints) and [OpenAPI import](/platform/mcp/open-api-import). #### Choose how customers authenticate If your tools need to know which customer is using them, choose how customers will authenticate when they connect from an AI client. | Method | When to use it | | ------------ | ---------------------------------------------------------------------------------------------------------------------------------- | | Customer key | Use customer keys when you want to create customer access through the API and provide each customer with their own MCP credential. | | OAuth | Use OAuth when customers already have accounts with your product and should continue signing in with their existing credentials. | With customer keys, you create a customer through the API. The customer receives an MCP URL and a customer key that identifies them when they connect from an AI client. With OAuth, the customer connects the MCP URL and signs in through your existing identity provider. Their existing account remains their identity when they use your product through an AI client. Both approaches allow agent activity to remain associated with the corresponding customer rather than creating a separate customer system for AI agents. Learn more in [Connect customer accounts](/guides/integration-guides/connect-customer-accounts). #### Connect from an AI client Once your MCP is configured, customers can add your MCP URL to a supported AI client. ```text https://mcp.yourdomain.com/mcp ``` Customers using a customer key connect with their MCP URL and customer key. Customers using OAuth add the MCP URL and complete your existing sign in flow. After the connection succeeds, the AI client can discover the tools you made available. A customer can then ask their AI agent to use your product in the same way they would normally use your application. For example, a customer using a commerce product could ask their agent to find a product and place an order. ```text Find a black running shoe under $150 and order my usual size. ``` The agent uses the descriptions and schemas you configured to determine which tools to call and what information each tool requires. #### Verify requests from Mitu When an AI agent uses one of your tools, Mitu sends the corresponding request to your API. Before processing the request, your backend should verify that the request came from Mitu. When you connect your API, you receive a signing secret that your backend uses to verify requests sent by Mitu. Keep this secret secure and only use it from your server. Request verification is separate from your customer authentication. It confirms that the request was sent through Mitu before your application trusts the request and processes it. If you use OAuth for customer authentication, the signed-in customer's access token is also forwarded with the request so your existing authorisation and account rules can continue to apply. See [Request verification](/platform/security/request-verification) for the verification process and implementation examples. #### Test your integration Connect the MCP from a supported AI client and confirm that the tools you configured are available. Ask the agent to perform an action that uses one of your tools. Your API should receive the request, verify that it came from Mitu, process it using your existing business logic and return its normal response. If the tools do not appear, confirm that the operations are enabled and that the customer has connected successfully. If a tool appears but the request does not reach your API, check your endpoint configuration and request verification logs. Once a tool can be discovered and called successfully, your product is ready to be used through supported AI clients. ## Your product is connected Your existing product can now be used from supported AI clients through MCP. Customers connect using a customer key or their existing account, and their AI agents can discover and use the operations you have made available. Your API continues to own your product and business logic. Mitu provides the agent-facing connection and sends requests to the API you already operate. From here, you can make additional operations available, accept payments from AI agents or issue Agent Wallets that your customers' agents can use to make payments on their behalf. ## Continue building #### [MCP endpoints](/platform/mcp/endpoints) Learn how your API operations become tools that AI agents can discover and use. #### [MCP authentication](/platform/mcp/authentication) Learn how customer keys and OAuth work when customers connect from AI clients. #### [Accept payments from agents](/guides/integration-guides/accept-payment-from-agents) Let AI agents pay while completing actions through your product. #### [Issue wallets to customers](/guides/integration-guides/issue-wallets-to-customers) Give your customers Agent Wallets that their AI agents can use to make payments.