Plans, offline use and lapses

Why plans count profiles and platforms, the 72-hour offline grace, and why dry runs stay free.

Plans, offline use and lapses

Plans count what costs: profiles and platforms. Each posting identity is someone's account acting, each platform a destination with its own rules and risk, so "how many identities, where" is the honest meter. Actions are never the meter: counting clicks would punish checking before publishing, which is the behavior the whole product exists to protect.

During early access there is one plan and it is free, with no tiers and no limits. When tiers arrive they will differ in profiles and platforms, and dry runs will stay free in all of them: a dry run publishes nothing, and refusing it would charge for caution. See Plans and limits.

The plan is checked at the moment of publishing, from an attestation your Layer5 Cloud organization issued after sign-in. Enforcement is off unless BLOWHORN_ENTITLEMENT=required is set in the environment on your Mac; it is a local switch, not an organization setting. With it on, a real run without a current attestation, or outside what the plan covers, refuses before any browser opens. The check reads no queue and no browser: the answer is already held, in the keychain, before the run starts.

With enforcement off, being offline changes nothing and nothing is ever refused for it. With it on, offline use degrades by the clock, not by features: a held attestation keeps publishing for 72 hours while Cloud is unreachable; past that, publishing refuses until the connection returns. Nothing else changes offline: reading the queue, dry runs, and reports never needed the plan and never ask for it. A lapse therefore reads as "not entitled, reconnect", never as lost work: the rows wait, the ledger waits, and the next connected run carries on where the last one stopped.

Last modified October 9, 2026: docs: sync from master (5f4e718)