The full corpus behind Klara MCP & API
The full corpus behind an MCP server and a REST API. Wire Klaracase into your assistant, your agent, or your own product.
Built for your tools
MCP for assistants like Claude and your agents, a REST API for your own applications. Both offer search, decision texts, the citation graph, datapackages and the citation check. Over REST, the citation check also reads uploaded PDF and DOCX files. Every answer comes with the decision it is grounded on.
~/.config/claude/mcp.json
{
"mcpServers": {
"klaracase": {
"url": "https://klaracase.de/mcp",
"headers": {
"Authorization": "Bearer <token>"
}
}
}
}# MCP over HTTP
curl -X POST "https://klaracase.de/mcp" \
-H "Authorization: Bearer <token>" \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call",
"params":{"name":"search_decisions",
"arguments":{"query":"Verkehrssicherungspflicht"}}}'# REST
curl "https://klaracase.de/api/v1/decisions/search?query=Verkehrssicherungspflicht" \
-H "Authorization: Bearer <token>"One call, a grounded answer
Every answer names its Aktenzeichen and paragraph number, matched against the corpus.
Metered by credits, authenticated by token
An API token belongs to exactly one organization. Every call, over MCP or REST, draws on that organization's credit balance, the same one the web app uses, at the same price. No separate billing, no separate account.