01
A stable engineering foundation
Requirements, context, constraints, feedback, testing, and deployment are made repeatable
OPCStack / Cloudflare nativeBuild OPC products faster, at lower cost, and ready for production
OPCStack is a Cloudflare-native AI SaaS template for One Person Companies. It includes authentication, payments, credits, D1 sharding, R2, KV, Queues, Cron, AI tasks, deployment automation, and a test suite
Runtime architecture Cloudflare Edge
01 / Worker runtime Live
SvelteKit SSR + Hono API Queue consumers Edge routes
02 / Control plane Primary
META_DB Shard registry User routing
03 / Tenant data Regional
D1 Shards 04 / Platform Attached
R2 · KV · QueuesOne Worker · one control plane · regional tenant data
Built for one person companies Open source template Cloudflare Worker runtime
Build Faster
AI can focus on product changes
Most templates provide pages, login, and a few sample APIs. You still have to build everything else from scratch02
Complete core capabilities
Stop rebuilding login, payments, storage, queues, and deployment for every project
03
A real product flow
Signup, beta codes, credits, referral rewards, payments, subscriptions, notifications, feedback, and Admin API are already wired
Production ready
More Production-Ready Than a Typical Cloudflare Template
∞ Near-unbounded horizontal growth
Edge Nearest Worker
Control plane META_DB
Tenant data D1 Shards
WNAMWEURAPACOC
Built-in
Identity and access
Better Auth sessions, email login, Google, GitHub and LinuxDO OAuth, beta-code gating and protected admin access
Built-in
Billing and credits
Subscriptions, payment webhooks, integer credit units, ledgers, grants, deductions and affiliate referrals
Built-in
Regional tenant data
A global META_DB resolves every user to regional D1 Shards with primary writes and read replicas
Built-in
Storage policy
R2 paths separate public, private and temporary objects with explicit ownership and lifecycle rules
Built-in
Async workloads
Queues, cron jobs and AI task state keep long-running work outside the request path
Built-in
Operations and delivery
Typed API contracts, generated migrations, local resources, tests and Cloudflare deployment automation
Why Cloudflare
Lower deployment and operations costs
Cloudflare puts Workers, D1, R2, KV, Queues, Cron, and CDN on one edge platform. It can run signup, usage, payments, async tasks, and file storage at low cost$5Workers Paid Plan starting price
0R2 egress charge
100KWorkers free requests / day
R2 egress is free, and KV, Queues, and D1 have no egress fees either
Quick Start
Create an OPCStack app with one prompt
- 01
Paste the prompt
Give the Agent the Quick Start link and your app name
- 02
Install the Skill
The Agent installs create-opcstack-app for its platform
- 03
Create the app
The Skill clones, configures, and starts the project
Agent prompt
Create an OPCStack app named <APP_NAME> by following:
https://raw.githubusercontent.com/glidea/opcstack/main/QUICK_START.mdFAQ
What to know before you start
Is this a template or a framework? Do I own the code after cloning
It is a template, not a framework. After cloning you own the full source, modules collaborate through explicit interfaces, you can change or delete any part, no framework conventions to follow and no vendor lock-in
Can I delete modules I don't use
Yes. Modules collaborate through interfaces, removing payment or AI does not affect auth or storage, you just skip writing the base flow instead of redesigning sign-up, payment and storage from zero
I'm not building an AI product, is it still a fit
Yes. AI is an optional module, remove it and auth, payment, credits, storage, queues and the deploy skeleton stay complete
How much can I actually save? Is the free tier enough to start
The free tier is enough to run an MVP, Workers Paid Plan is pay-as-you-go from $5/mo. In a public case a Vercel bill of $800+/mo dropped to under $20 on Cloudflare at the same traffic, and R2 egress is free
Once I've used the template, how do I sync later upstream updates
The repo keeps an upstream remote, follow SYNC_TEMPLATE.md to selectively merge upstream updates, it is not a one-time copy that cuts you off from the template
My project is small, is D1 sharding over-engineering
No impact. A small project runs on a single active shard by default, add shards when you scale, new users auto-assigned and existing users never migrate, sharding is a fallback for later, not a burden
What do I need to get started? Is init complicated
Just Node and pnpm. Give the Quick Start link to an Agent once to install create-opcstack-app, then use the Skill directly for future projects