MCP server #
The Kaddio MCP server lets AI assistants that support the Model Context Protocol (MCP) query and update Kaddio data through the GraphQL API. It uses the same tokens and access limits as the API.
The MCP server is in beta, and Kaddio support turns it on for your organization.
Connect a client #
Add Kaddio as a remote MCP server in your client, with:
- URL:
https://YOUR_ORG.kaddio.com/mcp - Header:
Authorizationset to your API token
Where these go depends on the client; see its documentation for adding a remote MCP server with a custom header.
Tools #
| Tool | What it does |
|---|---|
get_graphql_schema |
Returns the GraphQL schema, so the assistant knows what it can ask for |
graphql_query |
Runs a GraphQL query |
graphql_mutation |
Runs a GraphQL mutation |
List them yourself with:
curl -X POST https://YOUR_ORG.kaddio.com/mcp \
-H 'Authorization: YOUR_API_TOKEN' \
-H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Agent Skill for coding assistants #
The kaddio-graphql Agent Skill teaches Claude Code, Cursor, Codex and other coding assistants how to use the Kaddio API: schema discovery, authentication, rate limits, identifiers, webhooks and example operations.
The skill is available without a token at:
https://YOUR_ORG.kaddio.com/kaddio-graphql/SKILL.md
Install it with the skills CLI:
npx skills add https://YOUR_ORG.kaddio.com/kaddio-graphql/SKILL.md
Or save it in your assistant's skills folder, for example in Cursor:
mkdir -p .cursor/skills/kaddio-graphql
curl -o .cursor/skills/kaddio-graphql/SKILL.md https://YOUR_ORG.kaddio.com/kaddio-graphql/SKILL.md
Reading the skill needs no token, but running GraphQL requests still needs a token and the API add-on.