Connect a client
Point Claude or any MCP-capable client at this superlore deploy over Streamable HTTP.
superlore mounts its MCP over Streamable HTTP at:
/api/mcpOn this docs deploy it is public — no auth — so any MCP-capable client can connect.
Add the server
Point your client at the endpoint. For a Claude-style mcp.json:
{
"mcpServers": {
"superlore-docs": {
"type": "http",
"url": "https://YOUR-DEPLOY/api/mcp"
}
}
}Replace YOUR-DEPLOY with the host this KB is deployed at (for these docs, the site's own
origin). Then ask your agent things like:
- "search the superlore docs for the dual-representation contract"
- "get_page /docs/architecture"
- "list every node tagged canvas"
and it reads this knowledge base directly — through the six tools, not by scraping pages.
Inheriting auth
If a deploy gates the site (Auth.js / SSO — see Auth), the MCP inherits the same policy. This docs deploy is intentionally open so anyone can connect; your own KB can stay public, gate just the MCP, or gate everything.
Verify your own pages
After authoring, point an agent at the MCP and confirm search finds your page and get_page
returns the structure you intended. If the data isn't there, you wrote a picture, not a knowledge
face.