Two connected Instagram accounts is a real, if modest, ceiling — and it covers a genuinely common case: a main brand account plus a secondary one (a regional account, a founder's personal account, a sister brand). Whether one automation should serve both, or each needs its own, depends entirely on how similar their content actually is.
When sharing one workflow makes sense
If both accounts genuinely want the same underlying content — the same template, the same data, just published to two audiences — a single workflow with two Post nodes (each targeting a different connected account, fed from the same upstream Apply Template output) is simpler to maintain than two nearly-identical workflows. This fits regional variants of the same brand, or a primary and backup/archive account that's meant to mirror the main one.
When they need separate workflows
The moment the two accounts' voices, formats, or cadence genuinely diverge — a company account posting data-driven recaps and a founder's account posting personal commentary — forcing them through one automation means compromising one or both. A shared Custom Agent prompt tuned for one account's voice will read wrong on the other; a shared template built for one account's aesthetic won't fit the other's. In this case, two separate workflows (which can still share upstream data via the fan-out pattern if the underlying facts are genuinely the same) keep each account's content actually fitting that account.
A middle path: shared data, separate voice
The pattern that often fits best when two accounts are related but distinct: one HTTP Request node feeding two separate Custom Agent nodes (each prompted for that specific account's voice — see the Custom Agent prompting guide), each feeding its own Apply Template node and Post node. Same underlying fact, two genuinely different pieces of content, one shared and reliable data source underneath both.
What doesn't change between one account and two
Everything about the node mechanics — placeholders, bindings, triggers, the review habit before publishing — works identically whether a workflow targets one account or two. The only real decision is architectural: share the pipeline where the content is genuinely the same, split it the moment the accounts' actual voices diverge. Getting this call right up front saves a rebuild later, when a shared workflow starts producing content that visibly doesn't fit one of the two accounts it's serving.