DocsBuild with AI
MCP server
Connect Claude, ChatGPT or another AI assistant to your Kling account with the hosted Kling MCP server at api.kling.is/mcp, then ask it about your payments, revenue and subscriptions in Iceland.
Fastest: let your AI tool do the work
Copy the prompt and paste it into ChatGPT, Claude, Cursor, Lovable or any other AI tool. It reads our quickstart, asks about your business and picks the right setup.
The Kling MCP server connects an AI assistant such as Claude or ChatGPT to your Kling account, so you can ask it about your payments in Iceland and have it create payment links and checkouts for you, in Icelandic krónur. MCP (Model Context Protocol) is the open standard these assistants use to connect to other services. Kling hosts the server at https://api.kling.is/mcp, so there is nothing to install.
Connect your assistant
The connector lives at https://api.kling.is/mcp, and a Kling API key decides which account the assistant sees and what it may do there. Claude, ChatGPT and other assistants that add a connector from a link take your personal connector link: that address followed by your API key. You don't need to put it together yourself. The Connect AI assistants card on the API Keys & Docs page in the dashboard has the link for your test account ready to copy. In Claude, for example, you open the connector settings, add a custom connector and paste the link.
Tools for developers, such as Claude Code and Cursor, can instead connect to https://api.kling.is/mcp and send the API key as a bearer token, which keeps the key out of the link.
Treat the link like a password: anyone who has it can use your key. If it leaks, revoke the key in the dashboard and the link stops working.
What you can ask
Once the assistant is connected, ask in your own words. "What was my revenue last month?", "How has subscription revenue developed this year?", "Who are my biggest customers?" and "How many payments failed this week?" are all answered from your account's data. The assistant can also look up a single payment, a customer or a subscription, and list your products and webhooks.
It can create things too: "Create a payment link for 4.990 kr." gives you a link to send, and it can create checkouts, products, customers and webhook endpoints the same way. For developers, it can read the Kling API reference and write integration code against it.
Test or live, full or read-only
The key you connect with sets the limits. A test key only reaches test data, so it is the place to start. Once you take real payments you can connect a live key in the same way. If you only want answers, create a read-only key on the API keys page: the assistant then gets only the tools that read, and the API refuses any change it tries to make.
The tools
The read tools are get_checkout_session, get_payment_intent, list_payment_intents, list_products, list_customers, list_subscriptions, get_subscription, list_webhooks, the reports revenue_report, mrr_report, subscription_breakdown, subscription_movement_report, cohort_retention_report, payment_outcomes_report and top_customers_report, and get_api_docs, the Kling API reference. With a key that can write, the assistant also gets create_checkout_session, create_payment_link, create_product, create_customer and create_webhook.
On your own machine
The same server is also published as the npm package @klingis/mcp, for apps that run MCP servers on your own computer. It takes the same API key and offers the same tools. Most people don't need it: the hosted connector does the same with nothing to install.
To have an AI tool build Kling payments into your website or app, copy the prompt at the top of this page.