Whether to bring Postiz (a self-hostable social scheduler) into the stack for LinkedIn company-page publishing. My verdict is trial, not adopt. Three decisions are yours, and they are deliberately kept apart below so a boundary can't be nodded through inside a "yes".
The hard rule is "drafting only — no LinkedIn connection/DM automation, ever." Postiz does none of that. Every layer is publishing-only: the product ("does not automate or scrape content"), the LinkedIn Page API (only content/image/carousel fields — no connect, DM, or comment), and even its AI agent (commands limited to create-post, upload, analytics; "always a human in the loop"). Scheduled company-page publishing of your own content is the same category as your Postpone use — categorically distinct from the outreach automation the strategy rejects.
Trial, not adopt, because the LinkedIn path specifically has live friction two open issues confirm: LinkedIn's Marketing API version deprecates on a schedule (needs maintenance), and self-hosted LinkedIn OAuth has reported failures. A trial proves the path end-to-end before you rely on it.
| License | AGPL-3.0 — fine for internal self-hosting; obligations trigger only on redistribution, not our use |
| Health | 35.5k stars, active (~2,884 commits, ongoing releases) — not abandoned |
| Self-host | Docker; Postgres + Redis + Temporal; full feature parity with cloud — a real service stack to run and patch |
| Company-page posts via the official LinkedIn Marketing API; needs a LinkedIn Marketing Developer app (an approval step) | |
| Vs stack | Does not replace AuthoredUp (drafting) — adds scheduled publishing the stack lacks |
Every claim above is sourced — the working doc with each link is in the fleet tree (HI_T4_Postiz_Verdict_2026-09-05.md); this page carries the evidence so the decision can be made here.
Adopting a scheduler is a step from "drafting only" toward "publishing." It doesn't break the rule — but where the line sits is yours to draw, not mine, so I stopped at it rather than deciding it for you.
What I'd hold to: Postiz posts to the hi! company page and brand channels only. Your personal LinkedIn stays human — the strategy is explicit that "the pairing itself is content", and pointing a scheduler at your own profiles is the drift toward automation the plan deliberately refused. I've written that boundary into the deployment plan. Gate 2 below is you confirming it, or moving it.
Stand up a self-hosted Postiz and prove the LinkedIn company-page path end-to-end (draft → schedule → published post → analytics) before any real cadence runs. Not answering: the scheduling question stays open; no cost, no urgency.
Postiz is for the company page and brand channels only; founder personal LinkedIn stays human. Not answering: the trial shouldn't start without this line drawn — it's the whole reason a publishing tool doesn't become "LinkedIn automation".
Self-hosting means running and patching a Postgres + Redis + Temporal stack — real ops, heavier than the $20 AuthoredUp. Worth it only if multi-platform scheduled publishing is genuinely wanted. Not answering: could default us into ops work nobody chose.