Documentation

  • Getting started — a numbered walkthrough: the setup wizard, the home screen, connecting an agent, what it does, approving a card, the board page
  • Installation — npm, npx, from source, native module notes, upgrading, uninstalling
  • Using kanbo with agents — global setup, per project, per agent; the agent-vs-person rules
  • Orchestrator examples — an instruction file, Claude Code subagents, a Codex profile, a shell loop, a worked session
  • CLI reference — every command and flag, taken from kanbo <command> --help
  • MCP reference — every kanbo_* tool, its inputs and aliases, and what a list answers with
  • Configuration — every KANBO_* environment variable, the update check, .kanbo/binding.json
  • HTTP API — kanbo serve, routes, the security model, the board page
  • Storage — board files, shared Postgres/Supabase, roles, schema, migrations
  • Troubleshooting — kanbo doctor findings and error messages, Windows, WSL, Alpine, npx
  • FAQ — short answers to the questions people ask first
  • Performance — latency on boards of 100 to 10,000 cards, statements per read, MCP answer sizes, how to run the bench
  • Agent efficiency — the questions an agent asks, one call each: filters, pages, compact cards

See also the README, Contributing, Security and Changelog in the project root.