Guide7 min read
Faceless channels done right: how to automate informational videos
A channel where nobody appears on screen can be serious and useful, or a string of generic videos that no platform wants to recommend. The difference lies in the method, not the tool.
A faceless channel is a video channel in which the person who makes it does not appear: the story is told with a voice-over on top of images, stock clips, graphics or animation. In Spanish they are known as canales dark. Faceless YouTube channels are full of pieces mass-produced with AI, and the platforms have clarified and tightened their rules on that content. This guide explains what a faceless channel is, why so many fall by the wayside and how to set one up for informational videos with rigour, item by item and without promises.
What a faceless channel is, and what it is not
A faceless channel is a production format, not a business model. What defines it is that the person who creates it does not step in front of the camera. That leaves several ways of telling a story:
- Narration, human or synthetic, over stock clips or images.
- Animation, illustration or graphics that explain an idea.
- Screen recordings, maps or documents with commentary.
- Recreations with generated images or video, which have to be disclosed when they look real.
It is not a shortcut to making money without working. Courses sold on social media present these channels as a source of “passive income”. Be wary: in the United States, the Federal Trade Commission took action in 2024 against several businesses that promised that kind of income with “AI-powered” stores, as part of Operation AI Comply. A faceless channel is editorial work, with the camera switched off.
Why so many faceless channels fail
There is no single cause, but the same ones keep coming up. Almost all of them have to do with producing fast without any criteria for what to publish and what not to.
- Generic content. If anyone can produce the same video with the same tool, yours adds nothing. YouTube does not monetise content that looks as though it came off a template; we cover this in inauthentic content on YouTube.
- Factual errors. AI models can state false things with confidence. A made-up fact in an informational video can sink the credibility of the entire channel.
- Other people’s material with nothing added. Editing together other people’s clips with a narration that repeats what is on screen clashes with YouTube’s reused content policies and with Facebook’s policies on unoriginal content.
- Volume without review. YouTube’s spam policy does not allow automated tools to be used to flood the platform with similar videos made with minimal changes.
- Lack of transparency. Realistic AI-generated content has to be disclosed when it is published. Repeatedly failing to do so can lead to videos being removed or to suspension from the YouTube Partner Programme.
- Costs that nobody measures. Credits spent on failed attempts, overlapping subscriptions and review hours that nobody writes down. Without knowing what each published piece costs, you do not know whether the channel is sustainable.
- False expectations. Anyone who starts out expecting quick results gives up before the channel finds its voice.
How to run a faceless channel with rigour
Rigour does not depend on showing your face, but on how you choose, check and cover each topic. This is a six-step method that works with or without automation.
1. Define an editorial angle
Write down in one sentence which questions your channel answers, for whom, and what it will never do. “Business history told through its real figures” is an angle. “Fun facts” is not. The angle helps you rule out topics and gives every video a thesis.
2. Research with primary sources
Before writing, gather the sources: the body that publishes the statistic, the original document, the company’s report. Keep the verbatim excerpt that backs up each claim. If you cannot find it, the claim does not go in. You will find the full method in how to fact-check before publishing.
3. Write your own script
A good script does not summarise the source: it uses it to argue an idea. Add what the source does not say, such as a comparison, a consequence or some context. An AI model can help you draft it, but the thesis and the control of the facts are yours.
4. Choose voice and images that make sense
A synthetic voice can work if it is clear and consistent. Images should illustrate what is being said, not fill space. Use licensed clips, do not keep repeating the same ones and disclose realistic recreations made with AI. If you are unsure what to disclose, read how to label AI-generated content.
5. Review it as an editor would
Someone who did not write the script reviews all of it before it is published: facts, tone, image rights, captions and AI disclosure. Mechanical checks can be automated; the decision to publish cannot.
6. Publish at a sustainable pace and measure
Publish at the pace you can review. Measure which topics and formats hold attention, and what each published piece costs you. Correct yourself publicly when you get something wrong: it is a practice that the International Fact-Checking Network includes in its principles.
Tour
Rather see it in the dashboard?
A real recording of the CLAQON dashboard, with sample data, in under a minute.
Roles and costs, item by item
A faceless channel needs the same roles as any small media outlet. One person can cover several of them, but it is worth knowing what they are and which ones can be automated.
| Role | What it involves | Can it be automated? |
|---|---|---|
| Editorial direction | Choosing the angle, the topics and the thesis of each video | No: it is what sets your channel apart |
| Research | Finding sources and excerpts that back up each fact | Partly: the searching yes, the judgement no |
| Script | Writing the piece with a hook, a body and an ending | With help from an AI model and checks |
| Review | Checking facts, tone, rights and AI disclosure | The mechanical checks yes; the final decision no |
| Voice | Narrating the script | Yes, with a synthetic voice, or with your own voice |
| Editing | Putting together images, captions, music and effects | Yes, with varied styles |
| Publishing and analysis | Uploading, scheduling, measuring and learning | Yes, within each platform’s quotas and rules |
Costs break down into items that are worth recording separately. Prices change every month, so there are no figures here; we give you the list so you can draw up your own:
- People’s time, above all for editorial direction and review. It is usually the most overlooked item.
- Software licences, whether a cloud subscription or a licence to install on your own machine.
- AI providers: a text model for the script, and generated voice, images or video. Each one charges in its own unit.
- Search and research, if you use a search engine through an API.
- Stock clips and music, free or paid, with their licences.
- Hardware and electricity, if you produce on your own computer or server.
- Publishing: some platforms, such as X, charge for publishing through their API.
- Failed attempts: what you pay for pieces that never get published.
In how much an AI-made short costs you will find the formula for working out the cost per published piece with all these items.
Checklist before launching the channel
- You can explain in one sentence what your channel offers that no other does.
- You have a list of topics that fit the angle, enough for several months.
- You know where you will get the facts for each topic.
- Someone is responsible for reviewing each piece before it is published.
- You know which content you will disclose as AI-generated, and how.
- You have a licence for all the images, clips and music you are going to use.
- You know the cost per published piece, including failures.
- You know how you will correct a mistake if someone spots one.
How CLAQON handles it
CLAQON is designed for informational channels that want to produce at scale without giving up on sources. It does not promise audiences or income: it promises that every piece goes through the same checks and that you can see what it cost.
- Research and verification before the script. It searches for citable sources or uses the URLs you provide. The script can only rely on verbatim excerpts, and the deterministic HARD QC check rejects figures and absolutes that are not in the evidence. If a script does not pass after its attempts, it is held for human review before anything is spent on voice or images. More in source verification.
- Full production on your machine. Local voice with word-level timing (or ElevenLabs, if you enable it), licensed stock clips without repeating any from the last 45 days, editing with FFmpeg and publishing to YouTube, Instagram, Facebook and X. You can follow it step by step in the Shorts factory.
- Compliance controls. Synthetic content disclosure on YouTube when there are generated scenes, a similarity check against the account’s previous 30 pieces and a real-people guard on generated scenes.
- Spending with a cap. You pay each provider with your own keys, with no credits, and you set caps per video, per day and per month. We explain it in spend control.
You should also know what it does not do today. It does not publish to TikTok. It has a single user per installation. It renders videos on the CPU, one at a time, and only produces while your machine is switched on. And YouTube keeps videos uploaded through the API private until your Google project passes its audit. If your channel is informational and these conditions suit you, take a look at the informational channels use case.
CLAQON
Produce with sources and with spending under control
CLAQON researches, fact-checks, edits and publishes your informational shorts from your own machine. We are opening the early access programme to channel operators in Spain and Mexico.