Scheduled publishing: generate the audio at deadline, serve it at publication time
The newsroom wraps at night, the article goes out at dawn, and the embargo has to hold. How to schedule an audio publication without risking the 6am outage or an early leak, accurate to within two minutes.
There is a moment every fixed-schedule newsroom knows: the wrap happens in the evening, the article is ready, but it must not appear before dawn. The text version knows how to wait: content systems have scheduled publication for twenty years. Audio raises a question nobody asked while it did not exist. Do you generate it right away, at the risk of publishing too early? Or wait for publication time to run the synthesis, at the risk of the outage falling at exactly 6am, when no one is around to see it?
The real problem is not the synthesis, it is the calendar
Generating audio takes a few seconds, and that is not where it pinches. The sensitive point is elsewhere: the work happens when the newsroom is present (in the evening), and publication happens when it is not (at dawn, on a Sunday, on a public holiday). The two never fall at the same time, and it is precisely that gap a tool should absorb on your behalf.
Deferring generation to publication time turns every morning edition into a gamble: if the service is slow, if a queue forms, if an incident hits at 5:58, the article goes out with no audio and nobody notices before the day is under way. Generate right away and publish in the same breath, and you break the embargo: the audio of an embargoed article becomes a way around the embargo, and that is the kind of detail that gets noticed one time too many.
What scheduled publishing does
Scheduled publishing separates the two acts the calendar sets against each other. You set the publication time with the article, as a publish_at field (an ISO date with its timezone, for instance 2026-09-17T06:00:00+02:00). The audio is produced immediately, at deadline, when the load is absorbable and someone can still check it. But it is only served at the given time.
Precision is the point we verify and state plainly: the audio does not appear before the set time, and it appears within two minutes of it. Not "sometime in the morning", but on time, with a margin you can hold up against an embargo. Until the deadline is reached, scheduled audio is simply not served: nothing leaks, even though the file has existed since the evening before.
One case most systems handle badly remains, the last-minute fix. You reopen the article at 5:50 for breaking news. What becomes of the audio already generated? With us, the previous version keeps playing until the deadline, then the new one takes over: the player is never silent in between. It is the same concern for continuity as with regenerating a corrected article, where the principle is always to never leave a page with a button that plays nothing.
What it changes for a newsroom
The most concrete benefit is the one you never see: the load is absorbed at the right time. You produce in the evening, when you are there, and you guarantee audio available exactly at publication, when you are not. The morning edition no longer depends on the health of a service at the precise instant no one is watching it.
This mechanism combines with review. Approving audio in the evening does not publish it: it waits for its morning slot. A newsroom that requires an editorial green light before anything goes live can therefore approve the night before, calmly, without the audio leaving ahead of the article. The two logics (schedule and review) do not contradict each other, they stack.
Then there is measurement, because a feature that goes unmeasured eventually goes undefended. Once the audio is live on time, it is the listening figures that say whether it finds its audience, and what you should actually measure is worth pausing on before drawing conclusions: an on-time publication is worthless if no one then looks at what gets listened to.
Who it is for, and who it is not for
Let us be clear about the scope. Scheduled publishing answers a precise need: fixed publication times, embargoes, morning editions. A daily, a news site, an institutional title on a set rhythm will find it natural. A blog that publishes whenever an article is ready, with no imposed time, does not need it: immediate generation is enough, and adding a calendar constraint would only complicate a flow that asks for none.
It is a tool of editorial discipline, not a gimmick. It does not make the audio better, it makes it punctual, and for a newsroom that lives by the clock, punctuality is no detail: it is the difference between a complete morning edition and one missing its audio until someone reaches the office.
Give your articles a voice with WeDispatch
This blog is itself voiced by WeDispatch. Curious how it sounds on your content?
Book a demo