diff --git a/.planning/PROJECT.md b/.planning/PROJECT.md
new file mode 100644
index 0000000..50b930e
--- /dev/null
+++ b/.planning/PROJECT.md
@@ -0,0 +1,79 @@
+# SummerCMS
+
+## What This Is
+
+A Scala-based Content Management Framework (CMF), rewriting WinterCMS/OctoberCMS for better performance, scalability, and a cleaner developer experience. Targets the same audience as WinterCMS — agencies, freelancers, and developers who need to rapidly build anything from simple landing pages to complex portals.
+
+## Core Value
+
+A CMF that enables rapid application development with the flexibility to support any use case, delivered with the performance and horizontal scaling that PHP cannot provide.
+
+## Requirements
+
+### Validated
+
+(None yet — ship to validate)
+
+### Active
+
+- [ ] Model / Controller / Components system similar to WinterCMS
+- [ ] Modern frontend-backend communication (research needed — not plain AJAX, scalable)
+- [ ] Admin backend with YAML/JSON-driven form generation
+- [ ] CLI scaffolding commands for initial setup
+- [ ] Plugin system with interdependencies (plugins can extend each other)
+- [ ] Theme system with Vue compatibility
+- [ ] HTMX integration for data-attributes style interactions
+- [ ] Themes can declare rendering mode (SPA / hybrid / static)
+- [ ] Full WinterCMS feature parity (CMS pages, media manager, users, etc.)
+
+### Out of Scope
+
+- Multi-database support from start — PostgreSQL primary, others can come later
+- Snowboard/Alpine.js patterns — clean break from WinterCMS frontend mess
+- PHP compatibility layer — this is a rewrite, not a port
+
+## Context
+
+**Why this exists:**
+- OctoberCMS went paid license, development stalled despite thousands in yearly fees
+- WinterCMS has active development but messy frontend (Snowboard + Alpine.js + API mode + theme mode all mixed)
+- PHP performance and scaling limitations for high-traffic sites
+- Desire for cleaner, more maintainable codebase
+
+**Reference implementation:**
+- `golem15-wintercms-starter/` submodule contains WinterCMS reference
+- Patterns to replicate: plugin structure, YAML-driven forms, component system, event-driven model extensions
+- Apparatus framework concepts: scenario-based workflows, automatic DI, backend CSS/JS injection
+
+**Team:**
+- Principal Scala developer from major Scala company on board
+- Sufficient funding for AI agents and human developers
+
+**What made OctoberCMS great:**
+- Data attributes API for frontend interactions (HTMX captures this spirit)
+- Flexibility to build anything from landing pages to complex portals
+- Themes as first-class citizens
+
+## Constraints
+
+- **Language**: Scala with ZIO ecosystem — chosen for performance, type safety, and freedom-respecting community
+- **Database**: PostgreSQL primary — focus on one DB done well
+- **Frontend**: Vue-compatible themes + HTMX for interactions — clean break from Snowboard/Alpine mess
+- **Deployment**: Pulumi-first (IaC), Docker and JVM also supported
+- **Template engine**: TBD — needs research for Scala options
+- **Backend communication**: TBD — needs research (websockets scale poorly, explore alternatives)
+
+## Key Decisions
+
+| Decision | Rationale | Outcome |
+|----------|-----------|---------|
+| Scala over other languages | Principal dev expertise, performance, type safety | — Pending |
+| ZIO ecosystem over Typelevel | Community values alignment, opinionated framework fits CMF needs | — Pending |
+| PostgreSQL only for v1 | Focus beats breadth, can add others later | — Pending |
+| HTMX + Vue frontend | Cleaner than Snowboard/Alpine, captures OctoberCMS data-attributes spirit | — Pending |
+| Pulumi-first deployment | Modern IaC approach, cloud-native from start | — Pending |
+| Full WinterCMS parity as goal | Proven feature set, clear target | — Pending |
+| Incremental milestones | Validate architecture before building everything on it | — Pending |
+
+---
+*Last updated: 2025-02-04 after initialization*