Design archetypes, constrain templates and write AQL with an assistant that looks the answer up in CKM and the published specifications.
MIT-licensed and open source. Nothing to install to try it — point any MCP client at the hosted endpoint.
Archetypes come from CKM, type definitions from the BMM schemas, guidance from a corpus written against the published specifications. Nothing is recalled from training.
Search published archetypes and templates by concept, filter by RM class, and pull one back as ADL or JSON — in the editor you are already modelling in.
CKM, guides, examples, terminology and type specifications, reachable from Claude Code, Cursor, Claude Desktop, LibreChat or anything else that speaks the protocol.
The relevant implementation guide is read before the answer is written, so you get the convention and the anti-pattern — not just something that parses.
A knowledge and authoring server, not a clinical data repository. It reads specifications and published archetypes; it never stores patient data.
Use Cadasto's endpoint, or run your own instance over streamable HTTP or stdio.
The knowledge and tooling layer: tools, prompts, resources, and completions. Connect once from Claude Desktop, Cursor, LibreChat, or any MCP client.
The workflow layer for Claude Code and Cursor: skills, slash commands, and subagents that know when to load guides and call MCP tools.
Point any MCP client at Cadasto's endpoint:
{
"mcpServers": {
"openehr-assistant-mcp": {
"type": "streamable-http",
"url": "https://openehr-assistant-mcp.apps.cadasto.com/"
}
}
}