Set up Claude Code
One command. Signs in through your browser on first use.
The fastest route is still the paste: the agent runs these steps itself.
By hand
1. Add the SuperBooks server
Run this in any project. Claude Code stores the server in your user settings, so it is available everywhere.
claude mcp add --transport http superbooks https://api.superbooks.io/mcp2. Sign in
Inside Claude Code run /mcp, pick superbooks and choose Authenticate. A browser tab opens the SuperBooks consent screen; approve the scopes and come back. If the agent cannot open a browser, add the server with a key instead:
claude mcp add --transport http superbooks https://api.superbooks.io/mcp \
--header "Authorization: Bearer $SUPERBOOKS_API_KEY"Verify
`claude mcp list` shows superbooks as connected, and asking Claude Code to call the SuperBooks team_get tool returns your team's name.
What the agent can do
45 tools, gated by the credential's scopes: a Read Only key sees the read tier only, and destructive tools need two separate approvals. See Authentication and the API reference.
Claude Code signs in through the SuperBooks consent screen, so no key is copied anywhere. The raw instructions the agent executes are at /agent-setup/prompt.md; all guides are listed at Onboard your agent.