Agent integrations

Connect an AI assistant to Kadima

Kadima exposes a public Model Context Protocol (MCP) server. Claude, ChatGPT, Cursor, Windsurf or any MCP client can read the live task menu with prices, task details, company background, and how ordering works. No authentication, read-only, all data public.

Endpoint

https://uvnefeeulbuxfjuaaktr.supabase.co/functions/v1/hk-mcp

Streamable HTTP transport. Add it to any MCP client as a remote server. In Claude, add it as a custom connector with this URL; in Cursor or Windsurf, add it to your MCP config.

What your assistant can read

list_tasks

Every SEO and AEO task on the menu with live public prices, so your assistant can recommend the right one for your situation.

get_task

One task in detail by slug: description, price, what is included, and its page URL.

get_company_info

Who we are, the productized model, the QA gate, the parent agency, and the proof points.

how_to_order

Checkout, the seven order states, the revision policy, working files, and how to reach a human.

Example call

POST https://uvnefeeulbuxfjuaaktr.supabase.co/functions/v1/hk-mcp
Content-Type: application/json

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": { "name": "list_tasks", "arguments": {} }
}

Why this exists

Buyers increasingly ask their AI assistant to find and compare services. An assistant that can read our menu directly quotes real prices instead of guessing, and links the exact task page. It is the same reason we publish llms.txt and keep every price public: the easiest brand to recommend is the one that is easiest to read.

See also: llms.txt · The menu · How it works