Kumiko Documentation

What Kumiko is

A framework for config-driven, multi-tenant apps with realtime updates. Features declare themselves through defineFeature() — schema, commands, hooks, and screens in one place. Commands flow through a typed pipeline (validation → access → handler → audit → SSE broadcast), and state flows back to browsers over SSE.

Stack: Bun · Hono · PostgreSQL · Redis · React · Expo · bun:test.

What Kumiko solves

• Build internal tools fast without Retool/Internal lock-in — the code lives in your own repository.
• Multi-tenant apps without a six-month setup for tenant isolation, audit, and RBAC.
• Realtime out of the box — SSE plus optimistic locking is the default, not a premium tier.
• Schema-driven UI — one schema definition produces forms, lists, filters, and i18n labels. Designers and AI builders edit the same code.

What Kumiko is not

• Not Bubble or Webflow — code-first; the schema lives in your repo, not in a vendor’s visual builder.
• Not Retool or Internal — Kumiko owns the database; it does not decorate someone else’s.
• Not a headless CMS — Kumiko builds apps with business logic, not content pipelines.

Where to start

Section	Purpose
Quickstart	Get running locally in five minutes — Docker, dev server, first sample.
Walkthrough	Build your first feature end to end — schema, handler, UI.
Concepts	The mental model behind Kumiko. Read this before reading source.
Samples	Real apps and recipes you can run — asset tracker, helpdesk, PublicStatus.

How this documentation is organised

Section	Content
Get started	Local install, requirements, first-feature walkthrough.
Concepts	The mental model. Why Kumiko works the way it does.
Guides	Task-oriented how-tos: “I want to do X — show me.”
Bundled features	Pre-built features (auth, audit, files, …).
Platform	Deployment topologies (solo, Docker, k3s).
Reference	Auto-generated patterns, error codes, and sample READMEs.