Guides

This section is a placeholder. Nothing is published here yet.

Guides are the task-oriented pages: “add a subscription and meter it”, “authenticate a request through gateway”, “generate an SDK from a contract”. They assume you have read Getting started and know what you are building, and they get to the point fast.

Nothing here yet

A guide will land when a task is actually walkable end to end — which, in early development, means after the services it depends on exist. Writing a guide against commands that do not run yet produces the kind of confident, wrong documentation this site is trying to avoid.

What is walkable today, and where it lives

Those pages have not been written as guides because they are procedures rather than walkthroughs, and a procedure with a wrong step in it is worse than no guide at all. They are here instead, and every command in them was run against the real cafe binary and the real service repositories:

• Getting started — install cafe, validate a manifest, render and run a local stack, build its image. The whole path, with the five unbuilt commands marked.
• Upgrading — if you already run a deployment. What breaks in the contracts and the order to fix it in.
• Tenant provisioning — create an account, invite somebody, promote them, and verify it. Quoted from a live session.
• Backup and restore — what the backup covers, the bucket it will not save for you, the data you lose when a database is destroyed, and the restore that proves it works.
• Rotating secrets — Stripe, provider credentials, database passwords, object-storage keys, and the one that cannot be rotated.
• A service is down — which service, which kind of down, and what not to do.
• Billing webhooks failing — inspect, replay, recover, and why a replay is safe.
• Running the gates — the exact command for every repository, and the tiers that skip by default. Read this one before you believe a green badge.

In the meantime

• Contracts — how the services find each other. The one thing worth understanding before anything else.
• Architecture — what each service owns and why.
• Topology — ports, probes, environment variables, and the cross-repo drift audit.
• Observability — the telemetry contract, and the honest answer to “is the stack running”.
• Services — one page per service, each stating plainly what is built and what is not.
• the org — for what is landing right now.