YouTube Shorts Automation: Which Step Are You Actually Handing Over?
YouTube Shorts automation is not one thing you switch on. It is four separate jobs, and every tool on the market automates a different subset of them. Writing the script. Making the video. Deciding it is good enough. Putting it on your channel. A tool that does all four leaves you nothing to do and nothing to check; a tool that does the first two leaves you a publish button and a few minutes a day. Picking well means knowing which of those four steps is actually eating your week, and buying the tool that stops in the right place.
This article walks the four steps, then compares three tools that automate different spans of them: AutoShorts.ai, Revid AI and Nichecast. We will be upfront: Nichecast is ours, and it deliberately stops one step short of the others.
Everything below about AutoShorts.ai and Revid AI is quoted from their own pages as of September 2026: autoshorts.ai and revid.ai/pricing, with Revid’s figures re-read on 3 October 2026. Everything about Nichecast is what the product does today, checked against the running site. Prices and plan limits move; re-read the vendor’s page on the day you buy.
The four steps a Shorts channel actually runs
- Script. A topic becomes 100–150 words with a hook that survives the first second.
- Production. Voiceover, visuals, captions, music, a vertical render.
- Judgement. Somebody decides this particular video is worth your channel’s name.
- Publishing. The file gets onto YouTube with a title, a description and a time.
Every automation tool automates steps 1 and 2. The real difference is what it does with steps 3 and 4, and that is where the money, the risk and the daily time go.
What each tool automates
| AutoShorts.ai | Revid AI | Nichecast | |
|---|---|---|---|
| Script | Yes, from a Series topic or custom prompt | Yes, from text, a link or an idea | Yes, from your channel’s niche |
| Production | Yes | Yes | Yes |
| Judgement | Yours, if you preview before the schedule fires | Yours, unless Auto-Mode is running | Always yours: nothing goes public unreviewed |
| Publishing | Automatic, on the plan’s schedule | Automatic on paid plans, to TikTok, IG and YouTube | Never: the video waits as an unlisted draft |
| Unit you buy | A posting frequency | A monthly seat with an AI-credit allowance | One finished video |
| Editor | Preview and edit before posting | Full editor, avatars, face swaps, 100+ tools | None |
| Entry price | $19/month (Starter) | $39/month (Hobby) | $0 for 5 videos, then $29/month |
AutoShorts.ai: automation as a posting schedule
AutoShorts.ai’s home page states the pitch in one line:
Faceless Videos on Auto-Pilot. Our powerful AI video creation platform allows you to fully automate a faceless channel. Get views and grow while you sleep.
The unit of work is a Series: you pick a topic (preset or custom prompt), connect your channels, and the tool generates and posts on a cadence. Its own three-step explainer puts editing in the middle (“Review your AI-generated video before it’s posted”), so judgement is offered rather than removed, but it happens against a clock that fires whether or not you showed up.
That shape is priced accordingly. AutoShorts sells frequency, not videos:
| Plan | Price | What the page lists |
|---|---|---|
| Free | $0 | Creates 1 video, 1 Series, 0 motion credits, marked “Temporarily paused” |
| Starter | $19/month | Posts 3 times a week, 1 Series, 27 motion credits |
| Daily | $39/month | Posts once a day, 1 Series, 62 motion credits |
| Hardcore | $69/month | Posts twice a day, 1 Series, 124 motion credits |
Every tier lists the same feature set (auto-post, edit and preview, HD, background music, voice cloning, no watermark), so what you are moving up the ladder for is posts per week and motion credits, nothing else. There is a monthly/yearly toggle advertising two months free on annual billing.
Buy AutoShorts.ai if a fixed cadence is the point. If the reason your channel is stalled is that nothing goes out unless you sit down and do it, a tool that posts on Tuesday whether or not you are there is solving your real problem.
Revid AI: automation as a whole growth loop
Revid AI automates further out in both directions. Its pricing page describes the product as an end-to-end loop rather than a generator:
One membership. The whole growth loop. Create, publish, learn, repeat. Every plan includes the full engine: rendering, publishing, performance stats, Auto-Mode, and agent access.
The piece that matters for Shorts automation is Auto-Mode workers: background workers that keep producing, and they are metered by plan. As listed today:
| Plan | Price | Credits/month | Auto-Mode workers |
|---|---|---|---|
| Hobby | $39/month | Not stated on the plan | Included, number not stated |
| Growth | $99 struck through, $39/month, “Discount locked in forever” | 2,000 AI credits | 3 workers |
| Ultra | $199/month | 12,000 AI credits (selectable up to 200,000) | 10 workers |
Growth is also the tier where publishing to TikTok, Instagram and YouTube appears, along with post stats, 70+ language voiceovers, 4K output and API, MCP and CLI access. Ultra adds voice cloning and watching 50 channels for viral ideas.
Two things to hold onto. The Growth price is a promotion, not a list price. The page prints $99 struck through beside $39 and calls the discount “locked in forever”, which is a claim about the future only Revid can keep. And “2,000 AI credits” is not “2,000 videos”: the FAQ puts a typical 30-second video with Pro moving images and a voiceover at about 40 to 80 credits, roughly 25 to 50 videos a month, and says full AI video models cost more per second. That is a factor of two inside one estimate, so make one video of the kind you would publish and read what it debited before comparing it with anything. The same FAQ says every plan includes publishing and Auto-Mode workers, and that plans differ by monthly credits and the number of workers.
Buy Revid AI if you want the loop, not just the videos: create, publish everywhere, read the stats, and let Auto-Mode keep the queue full while you look at what worked.
Nichecast: automation that stops at the publish button
Nichecast automates steps 1 and 2 and then deliberately stops. You configure a channel’s niche once; it writes the script, generates the narration, the visuals and the captions, renders a vertical video, and uploads it to your own YouTube channel as an unlisted draft. It never publishes. You open YouTube Studio, watch the ones that landed, and press publish on the ones you stand behind.
That is one design decision with several consequences:
- You keep step 3 without doing step 2. The judgement stays human; the production does not.
- Nothing surprising ever appears on your channel. There is no schedule that fires while you are asleep, because publishing is not something the tool can do.
- You pay per finished video, not per cadence. The live pricing page lists two plans:
| Plan | Price | What you get |
|---|---|---|
| Free | $0 | 5 finished videos to start, once. The full pipeline, on your own channel, no card |
| Pro | $29/month | 30 videos a month, cancel any time from your account |
One finished video costs one credit, and so does drawing a character’s reference sheet. A video that fails anywhere in the pipeline refunds its credit automatically, so a bad run costs you review time rather than money. Cancelling stops the next charge and leaves the month you paid for running to its end.
The honest limit: there is no editor. You cannot nudge scene three. You accept the video or you generate another one. If shaping individual cuts is the part you enjoy, Nichecast removes the wrong thing, and Revid AI or a timeline tool will suit you better.
Rolling your own
The fourth option is a home-built pipeline: a script that calls a language model, a text-to-speech API, a renderer and the YouTube Data API on a cron. It is genuinely the cheapest per video, and people run real channels this way.
What it costs instead is maintenance you do not get to schedule: API changes, OAuth tokens that expire, render jobs that die halfway and leave a half-uploaded video, and a queue with nobody watching it. If you would enjoy maintaining it, build it. If you want it to be somebody else’s problem on the week your day job explodes, buy one of the above.
Choosing, in one pass
- Your problem is that nothing ever goes out → AutoShorts.ai. A schedule that fires without you is the whole product.
- You want distribution and analytics in the same place as generation → Revid AI, on Growth. Budget for the promotion ending.
- You want a continuous supply of finished videos but you want to be the one who publishes them → Nichecast. One credit per video, failures refunded, everything unlisted until you say otherwise.
- You want to own the pipeline → build it, and accept that you now run a small piece of infrastructure.
None of these is a quality ranking, and two of them can coexist. Plenty of channels generate in one tool and publish by hand anyway. The question worth answering before you pay anyone is the one at the top: which of the four steps is actually eating your week? Automate that one, and be suspicious of a tool that offers to take the step you would rather keep.
Related reads: we have compared each of these against Nichecast directly in An AutoShorts Alternative and A Revid AI Alternative, and looked at the wider category in AI Faceless Video Generator: Where Each One Stops.