INTEGRATION · STRIPE
Stripe is the source of truth for billing. We treat it like one.
A read-only restricted key, five required permissions, a live read check against each resource, and AES-256-GCM encryption at rest. No write access. No live keys.
How the connection works
You provide a restricted Stripe key. The backend verifies it can read customers, subscriptions, invoices, products, and prices before storing it. The key is encrypted with AES-256-GCM and a versioned prefix.
- Key must begin with rk_test_ in the current build
- Read permissions on five resources are required
- A read check runs against each resource before storage
- Live keys (sk_live_, rk_live_) are rejected
- The key is never returned to the browser after storage
What is synced
Stripe sync fetches subscriptions and invoices with pagination. Each sync is recorded as a complete snapshot with capture time and row count. The sync runs on demand or via cron-sync.
- Customer ID, subscription ID, status, plan, quantity
- Current period end and cancellation state
- Invoice data for dunning and payment context
- Paginated fetch with a 100-second timeout budget
Security boundaries
The platform key used for SubSync Guard’s own billing is separate from customer connection keys. Customer keys are encrypted at rest and never logged. Webhook verification uses the raw body and a signing secret.
Frequently asked
Can I use a live Stripe key?
No. The current build rejects any key that does not begin with sk_test_ or rk_test_. Live mode requires legal and payment approval.
Does SubSync Guard write to Stripe?
No. It only reads. It never creates, updates, or cancels subscriptions.
What if my key has too few permissions?
The connection is rejected. The backend checks read access on all five required resources before storing the key.