Distribute · faceless video channels
Pick an idea and Distribute writes the script, generates the scenes and the narration, and renders the episode. You approve the images and the final cut, from your phone if you like. Then it publishes on your schedule, with Shorts cut from every episode.
Setting up a channel is free. Every video shows its credit cost before anything is generated.
Launch your faceless channel
Start freeHow it works
The same five steps as the video. Each one is a screen you use in Distribute; the examples below are illustrations of those screens.
1Channel style
Look, image engine, narrator voice, on-screen captions, cadence and Shorts per episode are settings on the channel. Episodes inherit them, so the twentieth video looks and sounds like the first.
2Video Studio
Episode ideas queue up on one board. Before anything is generated you see the estimated credits, your balance and what is left after. Confirm, and the script, scenes and images build themselves.
3Scene review
Scene images are held for review before anything renders. Don't like one? Edit its prompt and regenerate it, or replace it. The render starts only when you approve.
4Mobile
The Wudtr app for iOS and Android lists every episode waiting for review. Watch the cut, approve or reject it, and an approved episode goes to its scheduled slot.
5Shorts
Each episode is cut into Shorts that open and close their own question, up to five per episode. They publish on their own daily rhythm to YouTube Shorts, TikTok, Instagram, X and Facebook Pages.
What comes out
From Wings & Wrecks, an aviation and maritime channel that runs on Distribute. Every frame, the voice, the captions and the thumbnail were produced by the pipeline above; a person approved the images and the cut. No camera, no editor.
Long-form
Why Rogue Waves Aren't Just Big Storms · a full episode, scripted from research, with animated diagrams
Short
A Short cut from a long-form episode, with its own hook and vertical frames
The player loads from YouTube only when you press play.
Your first channel takes a few minutes to set up and costs no credits.
Each episode moves through named stages. Two of them wait for you: the images, before anything renders, and the cut, before anything publishes.
Editorial rules
Structure settings live on the channel rather than on the episode. Change one and it applies to everything you make afterwards, so a lesson from your analytics only has to be applied once.
Open on the misconception, the story, the mechanism, the stakes or the practical use. Or compare, investigate, or run chronologically. Pick one per channel, or let the research pass choose per episode.
Production notes are injected into every script: your retention contract, the shape of the opening, and the phrases the channel must never use.
An explainer profile cuts fast; a documentary profile lets scenes breathe. Set the minimum and maximum runtime alongside it.
When the material will not carry the target runtime, the default is to ship a shorter honest video rather than pad it. You can override that per channel.
Synced storyboard mode gives each narration beat its own frame, cut on the real voice timestamps rather than an even split.
Cap how many frames one episode may generate, so a continuity-heavy channel stops producing near-duplicates it does not need.
Look and voice
Style, engine, voice and on-screen treatment are channel settings. Episodes inherit them, so the twentieth video looks like the first.
A written style prompt and a thumbnail directive hold the look steady, so every thumbnail shares one house style while the subject and hook stay per-episode.
Give the channel a named character with reference images so they appear consistently, plus a side cast for cameos. Or run it faceless with no host at all.
Seven engines, from the cheapest default to a premium tier: Gemini Flash and Pro, GPT Image 1 and 2, GLM, Magnific Mystic, or let it pick. Set image quality per channel and let weak frames be regenerated automatically.
Headline and caption, captions only, or images and voice alone. Five subtitle styles including karaoke and word-by-word, plus animated callouts on chosen keywords.
Six delivery styles, with speed, stability, similarity and expressiveness adjustable, over music from the built-in set or your own audio library.
Keep several intro and outro versions and the channel cycles through them, so a regular viewer is not shown the same twenty seconds every time. The outro can name the next episode automatically.
Six consecutive thumbnails from one channel. The subject and hook change per episode; the palette, framing and type do not.
Shorts
A segment has to earn its place. It should open a question a new viewer can follow, answer it inside the clip, and stick to one idea. Intros and summaries are excluded. If nothing in an episode qualifies, you get no clips from it, which is worth knowing.
Crop the parent for speed, regenerate the frames vertically so nothing is cut off, or add a purpose-written hook naming the part it covers.
Clips start at the strongest moment by default, not the opening — unless your channel deliberately writes a standalone short into its cold open.
Set the most shorts one episode may yield, up to five. A tight video gives three, a dense one gives more, and the limit only ever caps it.
Shorts publish on their own cadence with a set number per day, so a backlog drains steadily instead of six landing at once.
An opt-in playbook directs the hook, keeps visual beats short, and ends on a hard cut that loops — and suppresses the generic bumpers that do not belong in a short.
Post clips to YouTube Shorts, TikTok, Instagram, X and Facebook Pages, choosing per channel which ones apply.
Cadence, autonomy and organisation
There are three autonomy levels: you drive it, it runs and stops for your review, or it runs through on its own. Because that is a channel setting, one channel can be left alone while another stays fully manual.
Episodes per week with the days and hour you choose, a separate rhythm for shorts, and a spacing rule so clips from one episode do not land together.
When planned slots drop below your threshold, more are proposed. They wait for you to accept them, so the queue never publishes itself.
Every channel's episodes sit on one board by stage, so you can see what is waiting on you and what is waiting on a render.
Group episodes into series and file them under topics. A series also lets the outro of one episode name the next one.
Take a channel offline without deleting it, and duplicate a channel you have tuned to start the next one from a known-good setup.
Replies can be drafted within a daily cap and a time window, and held behind a review gate until you approve them.
Quality and insights
Two checks sit between render and publish, and settings changes are logged so you can read what happened after one against the change itself.
The finished video is scored and weak scenes can be re-rendered. Turn it off on a channel where you review every frame yourself.
Rendered but unpublished episodes are judged against the channel's own history for duplicate cuts, fragment titles and hook shape, before they reach an audience.
The two are plotted together, so a video that pulls clicks but loses people is distinguishable from one nobody opened.
Each settings change is recorded with what followed it, so a change that helped can be told apart from one that only coincided.
Each episode can produce a written article and a transcript into your content library, so the video is searchable as text too.
Each channel gets a public page and its own llms.txt, and an end-scene QR code can point viewers straight at it.
Distribute is in beta. Feature availability and limits may change while it is being built.
Output planner
Set a channel the way you would configure it and see the month it adds up to: uploads, credits, and the plan that covers them. This is arithmetic on your own settings and our published rates, not a benchmark.
Long-form 825–1,400 credits, Short 450–700. Shorts are capped, never quota-filled, so treat the counts as a ceiling. Optional AI video clips inside scenes are priced per second and not included here.
What it costs
There is no separate subscription for this module. Producing an episode consumes AI credits, so the plan you are on decides how many you can make in a month — and you can buy more at any point, including on the free plan.
A long-form episode costs roughly 825–1,400 credits end to end, and a short 450–700. The spread is real work: a longer script, more scene images and a longer voiceover all cost more, and the estimate for a specific job is shown before you commit to it.
The free plan includes no monthly credits, so producing on it means buying credits first. Everything that does not call a model — setting up a channel, its style and cadence, reviewing and publishing — costs nothing.
Questions
No. Channels are built from generated visuals and narration. You can give a channel a recurring host character with reference images so they look the same in every episode, but no camera and no voice of your own are needed.
Not unless you set the channel to do that. Autonomy is a per-channel setting with three levels: you drive it, it runs and stops for your review, or it runs through. By default the pipeline stops twice: on the scene images before rendering, and on the finished cut before publishing.
Distribute is in open beta on every plan, including Free. Producing a video uses AI credits: roughly 825 to 1,400 for a long-form episode and 450 to 700 for a Short, and the estimate for each video is shown before you confirm. The Free plan includes no monthly credits, so you buy credits before producing on it; paid plans include 1,000 to 13,000 a month.
Long-form goes to the YouTube channel you connect. Shorts can post to YouTube Shorts, TikTok, Instagram, X and Facebook Pages, chosen per channel, each through its own connected account.
As many as qualify, up to a ceiling you set (at most five). A segment has to open a question a new viewer can follow, close it inside the clip and carry one idea; intros and summaries are excluded. A tight video might give three. Zero is a possible answer, and it tells you the episode had no standalone moment.
Yes. Each channel has its own style, voice, cadence and connected accounts, and all of their episodes sit on one board by stage. You can duplicate a channel you have tuned to start the next one from a known-good setup.
By default the channel ships a shorter video rather than padding it to hit a target runtime. You can change that per channel if you would rather always hit the length.
Get started
Set up the channel, its style and its cadence for free. See the credit cost of each video before it is made, and approve every cut before it publishes.
Distribute is in open beta. Features and limits may change while it is being built.
