BeechCMS is a precision-engineered, solo-developer headless CMS built for developers who deliver high-performance websites for clients with zero infrastructure overhead.
The pitch is simple: you build the site, you hand it over, and your client can manage their own content forever — with zero hosting costs. No monthly subscriptions, no server bills, no maintenance contract needed just to keep the lights on.
This is possible because BeechCMS runs entirely on Cloudflare's free tier — Workers for the API, D1 (SQLite at the edge) for the database, and R2 for media storage. A Cloudflare account is all the infrastructure a client ever needs.
Warning
BeechCMS is currently in Beta and is undergoing its initial phases of review and case-study testing.
- Known Risks: The software contains known security vulnerabilities, risks of data loss, and is completely untested in production environments.
- Liability: We accept absolutely no responsibility or liability for any damages or data loss. The software has not yet reached the security and stability standards required for production use.
- Intended Use: Use is strictly recommended for evaluation and MVP (Minimum Viable Product) testing purposes only.
CRITICAL REQUIREMENT: If you still choose to deploy or use BeechCMS, you MUST keep the software constantly updated to the latest commit/release to mitigate emerging risks.
Cloudflare's free tier covers D1, R2, and Workers for the vast majority of real-world content sites. You deliver a fully self-managed product — dashboard, API, media uploads — and the client pays nothing to keep it running. That's a compelling offer.
Unlike other "flexible" CMSs that store data in slow JSON blobs, Beech generates dedicated SQL tables for every content type. Enjoy native B-Tree indexing, REAL and INTEGER types for mathematical operations, and ultra-fast queries.
Define your content model dynamically via the dashboard UI or in code (TypeScript). The Botanical Engine compiles your definitions into deterministic SQL DDL at runtime. Bootstrap your schema from code using beech onboard or beech seed:load, then let the D1 database act as the single source of truth—with real-time updates and zero server downtime.
Your project is four files. The BeechCMS engine, dashboard, and API live inside node_modules/@beechcms/api. Update with pnpm update @beechcms/api.
npx @beechcms/cmsThe interactive wizard scaffolds a ready-to-use project in seconds. For everything from configuration to deployment, see the Developer Guide.
The heart of BeechCMS is the Botanical Engine, a high-performance Schema Compiler that bridges the gap between TypeScript definitions and SQL infrastructure.
Instead of generic document storage, the Engine analyzes your Seeds (content types) and compiles them into dedicated SQL tables within Cloudflare D1. Every field you define becomes a native, type-safe SQL column, allowing for:
- Native Performance: Real B-Tree indices and FTS5 virtual tables for ultra-fast filtering and full-text search.
- Data Integrity: Native SQL types (REAL, INTEGER, TEXT) with CHECK constraints for robust data handling.
- Zero-Manual SQL: Deterministic DDL generation—defined in code or via the UI, and the Engine handles the database synchronization at runtime.
| Layer | Technology |
|---|---|
| Edge runtime | Cloudflare Workers (V8 Isolates) |
| API framework | Hono v4 |
| Database | Cloudflare D1 (SQLite at the edge) |
| Object storage | Cloudflare R2 (zero egress cost) |
| Shared logic | @beechcms/core — Botanical Engine, types, validation |
| Dashboard | React 19 + Vite 7 |
| UI | Tailwind CSS v4 + Shadcn/ui |
| Server state | TanStack Query v5 |
| Rich text | TipTap v3 + KaTeX |
| Auth | jose JWT + bcryptjs |
| Validation | Zod v4 |
| Testing | Vitest v3 |
| Build orchestration | Turborepo v2 |
| Document | Description |
|---|---|
| Developer Guide | End-to-end guide: scaffolding, seeds, API consumption, deployment |
| Architecture | Monorepo topology, Botanical Engine data flow, D1 model |
| API Reference | Auth, Content CRUD, Media Engine, Public API, rate limiting |
| Frontend Guide | FieldRenderer registry, TanStack Query patterns, adding field types |
| System Map | Folder structure, conventions, and architectural constraints |
One afternoon I was staring at half a CMS with no name. I'd spent weeks building my desk setup — monitors, peripherals, the works — and the last thing I added was a beech wood desktop. I cut it, shaped it, sanded it, finished it until it was exactly right.
I was sitting at that desk trying to think of a name, looked to the side, and saw the wood. Beech. And then it clicked — seeds, branches, the forest, the fruit. The whole metaphor was already there in the project: content types are seeds, fields are branches, the data that grows from them is the fruit. The name took about three seconds.
Sometimes the best names aren't invented. They're found.
In 2025 I lost a €2,000 contract for a React showcase site for a construction company. The client needed two things beyond a static site: a blog where they could post photos of new builds, and a contact form. Simple requests — but they turned the project into a problem.
The tools that existed were either too heavy (WordPress, which the client had already had a bad experience with), too expensive once you added hosting and a managed database, or simply not designed to work alongside a modern React frontend without standing up a dedicated backend. The monthly running costs pushed my quote above what a larger agency — one that already had its own internal tooling — could offer. I lost the contract.
That was the moment I understood the real gap. Small agencies and freelancers don't lose on talent or quality. They lose because they don't have the same leverage as larger competitors who've already amortized the cost of building their own CMS. I decided to build mine.
Why Cloudflare? I was already looking for an alternative to Vercel that didn't require a paid plan for commercial projects. When I discovered that Cloudflare's free tier included not just Workers but also D1 (a relational database at the edge) and R2 (S3-compatible object storage), the whole architecture became clear. A full CMS backend — API, database, media storage, auth — with zero monthly cost for the client.
Why the Botanical Engine? I built the Botanical Engine as a high-performance Schema Compiler. It bridges the gap between the flexibility of a CMS and the power of a relational database. Instead of storing data in generic blobs, it compiles TypeScript definitions into native SQL tables and columns. This ensures rigid data integrity (crucial for things like financial records) and unlocks massive performance gains by leveraging native D1 indexing. This true relational model maintains the simplicity of "Schema-as-Code" while allowing for advanced features like cross-seed relations and complex aggregations.
What I learned building it. BeechCMS is my first serious serverless project. I had no prior experience with Cloudflare Workers, edge computing, or D1. Everything I know about this stack I learned by building this — reading documentation, hitting limits, understanding why they exist, and finding the right abstractions. The project is currently in active development toward a public 1.0 release and is already available on pnpm.
BeechCMS — Precision-engineered content infrastructure for the edge.
| Package | License | Details |
|---|---|---|
@beechcms/core |
MIT | Free for any use, including commercial |
apps/api |
BUSL-1.1 | Free for self-hosting; see Additional Use Grant |
apps/dashboard |
BUSL-1.1 | Free for self-hosting; see Additional Use Grant |
Self-hosting is always free.
The BUSL-1.1 only restricts offering BeechCMS as a managed SaaS service
in direct competition with BeechCMS Cloud.
After 2030-05-26, these components convert automatically to GPL v2+.
For commercial licensing inquiries: demusso1617@gmail.com
For more detailed information, see our Licensing Guide.