To build a Salesive App you register it, implement OAuth 2.1 with PKCE, request the scopes you need, handle signed webhooks, test with review testers, and submit it for review before it appears in the marketplace. Development and production configuration share one client ID, so the app you test is the app you publish.
What do you need before you start?
You need three things: a service you can host and keep online, a clear job the app performs, and a Salesive store to test against. The job matters most. An app that tries to do everything for every merchant will struggle in review and confuse the people installing it. Pick one problem, such as syncing orders to an accounting tool, and do it well.
Decide early which parts of the store you actually need. Write the list of resources the app touches, then map each one to a scope. That list becomes your consent screen, your review notes and your security boundary. If a resource is not on the list, the app should not be asking for it.
Why does development and production share one client ID?
Salesive gives an app one client ID for both development and production. There is no second registration and no separate identity to keep in sync. The app that a review tester installs is the same app that goes live, which removes a whole class of mistakes where the tested build and the published build are not quite the same.
Your own environments should still be separate. Keep development and production configuration apart on your side, with different databases and different secrets, even though the client ID is shared. The platform's identity is stable; your deployment should still be deliberate about which environment it is running in.
How does the OAuth flow work?
The merchant starts the install from your app, is sent to Salesive, signs in, reviews the scopes and approves. Salesive returns the merchant to your redirect and issues a token for that store. The flow uses OAuth 2.1 with PKCE, which protects the exchange even though a web app cannot keep a client secret perfectly.
Store the token per store, not per user and not globally. An app installed on two stores should hold two tokens, and every API call should use the token for the store being worked on. If a merchant revokes the install, that token stops working. Do not build a fallback that keeps a stale token alive.
Which scopes should you request?
Request the smallest set that does the job. If the app reads orders and writes nothing, ask for read access to orders and stop there. If it manages discounts, ask for discounts and not for customers. The scopes cover orders, products, customers, categories, shipping, discounts and blogs, and each one is a separate decision for the merchant.
A long permission list is not a sign of a powerful app. It is a sign that the merchant has to trust you with more than the job requires, and that makes the consent screen harder to justify. Narrow scopes also make the app easier to review, because the reviewer can see why each permission exists.
How do webhooks fit in?
Use signed webhooks for events rather than polling the API on a timer. Salesive signs the request, and your app verifies the signature before acting on it. That check is not optional: without it, anyone who learns your endpoint URL can send fake events and trigger whatever your app does in response.
Handle duplicates. A webhook can arrive more than once, and a retry after a timeout is normal. Make the handler idempotent by recording the event id and ignoring one you have already processed. Also respond quickly and do the slow work afterwards, so a slow handler does not cause a queue of retries.
Log the failures as well as the successes. When a webhook is rejected or an API call returns an error, record the store, the event and the response so you can replay the case later. A silent failure in an app looks like a missing order to the merchant, and that is the hardest kind of bug to explain.
How do you test and publish?
Review testers can install an app while it is still in review. Use that. Install it on a real store, walk through the consent screen, revoke it, reinstall it, and check that the token behaves the way you expect. Test the webhook signature with a deliberately wrong signature and confirm that your app rejects it.
Then submit for review and publish to the Apps marketplace once it is approved. Before you submit, make sure the listing explains what the app does, which scopes it needs and why. A reviewer who has to guess will ask questions, and a merchant who has to guess will not install. If you plan to charge, read how app billing works before you launch.
A good app is not the one with the most permissions. It is the one that does a clear job with the smallest key.
Frequently asked questions
Do I need a separate app for development and production?
No. Development and production configuration share one client ID. Keep your own environments separate, but register the app once. The identity follows the app from testing to live.
Can I publish an app without review?
No. Apps are published to the marketplace after review. Review testers can install the app while it is still in review, so you can test the real install flow before the listing is public.
What can review testers do?
They can install the app while it is in review. That lets you exercise the consent screen, the token and the webhook flow against a real store before the app is listed publicly.
Do I have to support billing?
No. Billing is optional. If you want to charge for the app, Salesive handles the approval and the payout, and billing must be enabled for your payout store first. The 5 percent fee and the charge flow are covered in the billing guide.
Which scopes will my app need?
Only the ones that match its job. Orders, products, customers, categories, shipping, discounts and blogs are granted separately. Start from the smallest set and add a scope only when a feature genuinely cannot work without it.
Building an app is mostly a discipline problem: one clear job, the smallest scopes, a token per store, verified webhooks and a real install test before review. Do that and the review is a checkpoint rather than a wall. Start with what Salesive apps are, then read app billing if you plan to charge. For the platform, see features and pricing, or browse the Salesive blog.