AI Native Visual Engine

Implementation-oriented specification repository for an AI-native visual editor engine.

Contents

• docs/spec/README.md: overview and design principles
• docs/spec/domain-model.md: document, page, scene, node, layout, and state ownership
• docs/spec/runtime-engine.md: runtime actions, command bus, middleware, handlers, and history
• docs/spec/semantic-system.md: semantic actions, compiler pipeline, and AI schema
• docs/spec/plugin-system.md: plugin contracts, registry, and renderer boundaries
• docs/spec/engine-api.md: public API surface for plugins, components, and the editor shell
• docs/spec/component-types.md: built-in types and default plugin component definitions
• docs/spec/component-states.md: named component states, props overrides, and exclusive state groups
• docs/spec/prototype-components.md: user-definable reusable component presets and instance inheritance
• docs/spec/rich-text.md: rich text model, Tiptap integration, and Markdown interop
• docs/spec/editor-interaction.md: selection, transform, drag and drop, multi-page, and collaboration
• docs/spec/collaboration-framework.md: Yjs transport, action sync, presence, and conflict convergence
• docs/spec/collaboration-infrastructure.md: Cloudflare Workers + Durable Objects relay and persistence
• docs/spec/history-and-undo-redo.md: undo/redo model, inverse-action contracts, and collaboration-history interaction
• docs/spec/constraints-and-validation.md: constraint system and validation layers
• docs/spec/schema-validation.md: Zod schema strategy, trust boundaries, and plugin schema contract
• docs/spec/roadmap.md: phased delivery and acceptance criteria
• docs/spec/document-runtime.md: document-level actions, command bus, history, and event log
• docs/spec/bootstrap-and-lifecycle.md: document initialization, session lifecycle, and import/export
• docs/spec/renderer-contract.md: renderer interface, rendering modes, and fallback policies
• docs/spec/testing-and-fixtures.md: test strategy, fixture format, and per-phase test requirements
• docs/spec/persistence-and-serialization.md: storage, serialization, snapshot, and event log truncation
• docs/spec/error-handling.md: error taxonomy, propagation, and recovery strategies
• docs/spec/theme-and-tokens.md: theme model, token system, and cascade rules
• docs/spec/data-binding.md: data source binding, variable resolution, and binding lifecycle
• docs/spec/data-interaction.md: cross-filtering, drill-down, drill-through, and filter integration
• docs/spec/implementation-stack.md: engineering toolchain, package layout, and runtime environment
• docs/spec/deployment.md: Cloudflare Pages deployment and CI pipeline
• docs/adr/README.md: accepted architecture decisions and ADR backlog

Goal

Define the architecture and delivery plan for an engine that supports:

• AI page generation
• AI-driven page updates
• scene graph based editing
• multi-page dashboards
• runtime rendering
• plugin extension
• undo and redo
• collaboration

Status

Phases 1-6 complete. Core engine, editor interactions, semantic compiler, AI integration, collaboration, and deployment infrastructure all implemented.

Local Development

# Install dependencies
pnpm install

# Build all packages
pnpm build

# Run tests
pnpm test

# Lint and format
pnpm exec biome check --write

# Start editor dev server (requires packages/editor setup)
cd packages/editor && pnpm dev

Collaboration (local)

# Start collab relay worker locally
cd packages/collab-worker && pnpm dev

CI/CD

• pnpm exec biome check — lint + format, enforced in CI
• pnpm -r exec tsc --noEmit — typecheck all packages
• pnpm -r test — run all tests
• PR previews deployed to Cloudflare Pages automatically