Budibase

# MISSION: Scaffold "Budibase" Production Architecture
You are an expert Senior Staff Software Architect and Full-Stack Engineer. Your mission is to scaffold and implement a production-grade, highly reliable, and modular codebase following the proven architecture of **Budibase**.
---
## 1. PROJECT SPECIFICATIONS & BENCHMARK
- **Reference Architecture**: Budibase
- **What It Does**: Budibase is an open-source, self-hostable platform for building AI agents, automations and internal business apps from a visual builder, backed by a Svelte builder/runtime, a Koa API and worker layer, and a CouchDB system database.
- **Domain & Category**: Internal Tool Builder
- **Production Scale**: 6-20 people
- **Development Mode**: CLASSIC
- **Architectural Rationale**: Splits the builder (Svelte), API/automation engine and background jobs (Koa) into independently deployable packages on shared CouchDB, while isolated-vm sandboxes user scripts and AI tool calls.
- **Live Website Reference**: https://budibase.com
- **Source Repository**: https://github.com/Budibase/budibase
---
## 2. PRODUCTION TECH STACK
- **Full Stack Array**: TypeScript, Svelte, Koa, CouchDB, Redis, Docker, Kubernetes, Vite
- **Primary Language(s)**: TypeScript, Svelte, JavaScript
- **License of the reference repo**: NOASSERTION
- **Frontend**: Svelte — Both the app builder (packages/builder) and the published app runtime (packages/client) are Svelte; bbui is the shared Svelte component library.; Vite — Bundles bbui, client and the frontend packages; builder combines it with routify for file-based routing.; CodeMirror — Powers the builder's binding, JavaScript and JSON editors (codemirror plus @codemirror/lang-* packages).
- **Backend & APIs**: Koa — Both packages/server and packages/worker are Koa applications, using @koa/router, koa-session and koa-passport for auth.; Node.js — Runtime for server, worker and CLI packages; version pinned for local development.; isolated-vm — Sandboxes JavaScript execution for automation bindings and scripting steps outside the main process.; Bull — Backs scheduled automations and background jobs in backend-core and worker, queued on Redis.
- **Data & persistence**: CouchDB — Primary datastore for app definitions, rows and metadata; the hosting image is a customized CouchDB build.; Redis — Caching, session storage and distributed locks via ioredis and redlock in backend-core.
- **Infrastructure & deploy**: Docker — hosting/docker-compose.yaml composes the server, worker, CouchDB, proxy and Redis containers for self-hosted installs.; Kubernetes — A Helm chart under charts/budibase, with chart-testing config, supports cluster deployments.
- **Tooling, testing & ops**: Vercel AI SDK — Powers the agent chat and tool-calling surfaces (ai plus @ai-sdk/svelte) that frontend-core exposes to the builder.; LiteLLM — Self-hosted proxy configuration (litellm_config.yaml) that normalizes calls to different AI model providers for Budibase agents.
---
## 3. KEY ARCHITECTURAL DECISIONS (audited from https://github.com/Budibase/budibase @ c024aad)
1. **The product pivoted from "low-code apps" to "AI agents, automations and apps"**: The README and package topics (ai-app-builder, ai-applications alongside low-code, crud-app) describe Budibase as an operations platform for agents, automations and apps; the frontend-core package pulls in the Vercel AI SDK and hosting/ ships a LiteLLM proxy config, layered on top of the original low-code app builder.
2. **CouchDB is the system of record; external SQL/NoSQL sources are a separate layer**: backend-core depends on @budibase/nano and pouchdb/pouchdb-find to talk to the bundled CouchDB, which stores app definitions, users and table rows. Connecting a user's own PostgreSQL, MySQL or other database is a distinct 'datasource' integration, not the system database.
3. **Server, worker and builder are separate deployable packages sharing typed core libraries**: packages/server (the API and automation engine) and packages/worker (background jobs, global config) are independent Koa apps with their own Dockerfiles, both depending on packages/backend-core, packages/shared-core and packages/types for shared logic and typings, so the two processes can be scaled independently.
4. **Automation and binding code runs inside isolated-vm, not the main process**: packages/worker and packages/server depend on isolated-vm (^6.0.1) to execute user-authored JavaScript (automation steps, handlebars helpers via string-templates) in a V8 isolate rather than Node's main context.
5. **Nx and Lerna both configure the same yarn workspace**: nx.json defines the task graph and caching for the monorepo, while lerna.json is still present for legacy versioning/publish flows; scripts like `build` invoke `lerna run build --stream` rather than a bare `nx` command.
---
## 4. NON-NEGOTIABLE ARCHITECTURAL GUARDRAILS
1. **Monorepo & Modular Separation**:
   - Structure as a Turborepo monorepo with strict package boundaries:
     - `apps/web`: Application UI, routing, layouts, and server endpoints.
     - `packages/ui`: Shared design tokens, CSS variables, and Radix UI primitive components.
     - `packages/db`: Database schemas, client singleton, declarative migrations, and seed scripts.
     - `packages/config`: Shared TypeScript, ESLint, and build configurations.
2. **Strict Type Safety & Zero `any` Policy**:
   - Enable `strict: true`, `noImplicitAny: true`, and `strictNullChecks: true`.
   - Validate ALL external inputs, API request bodies, and query parameters with **Zod** schemas before execution.
3. **Frontend & Rendering Guidelines**:
   - Utilize Vite 6 with TanStack Router for fully type-safe routing. Manage server state and caching via TanStack Query v5 with optimistic UI updates.
4. **Design System & Aesthetics**:
   - Keep every color, radius, shadow and font in a single token file (CSS variables) and consume tokens everywhere; never hardcode hex values in components.
   - Prefer crisp 1px borders and one subtle shadow scale over blurry default shadows. Pair one sans-serif for body/headings with one monospace for tags, badges, metrics, and code.
5. **Data Layer & Reliability**:
   - Write declarative schema definitions with foreign keys, composite indexes on queried filters, and automated timestamp triggers.
   - Use connection pooling and prepared statements for serverless database execution.
---
## 5. STEP-BY-STEP SCAFFOLDING ROADMAP
- **Phase 1: Workspace & Root Config**: Initialize package manager, monorepo configuration (`turbo.json`, `tsconfig.base.json`, `package.json`).
- **Phase 2: Database Schema & Client**: Set up the data layer (CouchDB, Redis): client, connection pool, models, and migration scripts.
- **Phase 3: Design Tokens & UI Primitives**: Build accessible `Button`, `Input`, `Card`, `Badge`, and layout wrappers inside `packages/ui`.
- **Phase 4: Core Application Routes & Handlers**: Implement primary authentication, user session handling, and application routes.
- **Phase 5: Quality Assurance & Build Verification**: Run `tsc --noEmit`, ESLint, Prettier, and smoke test suites to ensure zero compilation or runtime errors.
---
## 6. EXECUTION INSTRUCTIONS
1. Review all specifications, architectural guardrails, and stack choices above.
2. Present the full monorepo directory tree structure.
3. Systematically generate the complete, production-ready codebase according to the 5-phase roadmap above — starting with the root workspace setup, followed by the database schema, UI design system package, and full-stack application routes until the repository is fully scaffolded and ready to run.