How long does a faceless video take to make?
The tooling steps take minutes and cost cents; the human steps take about seven hours and cost nothing, which is why people budget for the wrong one.
Disclosure: some links on this page earn us a commission if you subscribe, at no extra cost to you. It never changes what we recommend — where a free or cheaper tool wins, we say so.
Faceless video is sold on the premise that automation does the work. The tooling genuinely is fast — narration renders in seconds for about three cents, and capture software is free.
Which is precisely the trap. The steps that got automated were never the expensive ones.
The two kinds of time
Split every step by what it consumes:
Machine time — you start it and walk away. Rendering narration, encoding, uploading. Costs money, not attention. Parallelisable. Already close to zero.
Human time — it takes exactly as long as you are sitting there. Deciding what the video is about, checking whether the claim is true, writing 1,200 words that are worth hearing, judging whether a cut works.
Automation has crushed the first category and barely touched the second. Any plan built on “the AI makes the video” is budgeting the wrong column.
Where an 8-minute video’s time goes
Using this site’s own figures: an 8-minute video needs about 1,200 words of script.
| Step | Type | Rough time | Cost |
|---|---|---|---|
| Pick the topic, confirm it has substance | human | 15–45 min | — |
| Research and verify the claims | human | 1–3 hours | — |
| Write 1,200 words for the ear | human | 1–3 hours | — |
| Render narration | machine | seconds | ~$0.03 |
| Listen back, fix what does not land | human | 20–40 min | pennies to re-render |
| Capture / assemble visuals | human | 1–3 hours | $0 |
| Edit, caption, tighten | human | 1–3 hours | $0 |
| Thumbnail and title | human | 20–60 min | — |
| Upload, metadata, disclosure | mixed | 15 min | — |
Estimates, not measurements. We have not yet produced at volume, so these are a model to plug your own numbers into rather than times clocked from our own production. The ratio is the durable part; the absolute figures will vary enormously by topic and experience.
Totals land somewhere around 5 to 12 hours per video — against roughly thirty cents of machine cost.
The ratio is the finding
Machine time is under 1% of the total. Cost per video is under a dollar. Human time is essentially all of it.
Two things follow, and they are the whole point of this page:
Tool-shopping is procrastination. Moving from a free voice to a premium one changes a step that already takes seconds. It does not touch the seven hours around it. The hours you spend comparing tools would have been better spent producing.
The economics are a wage question, not a cost question. Monthly running costs are genuinely small, which is real and worth knowing. But at 5–12 hours a video, the binding constraint is your time. A channel publishing weekly is a serious part-time job, and pretending otherwise is how people arrive at video nine, exhausted, having been told this was passive.
What actually gets faster
Not by buying things. By these, roughly in order of effect:
- Picking topics you already understand. Research is the single largest block, and it collapses when you know the subject. This is an argument for a narrow niche that has nothing to do with SEO.
- Reusing a format. Not a template — YouTube’s inauthentic content policy is aimed exactly at interchangeable output. But a consistent structure, with genuinely different substance each time, removes the blank-page cost without triggering it.
- Writing for the ear on the first pass. Rewriting read-prose into spoken prose is a second full pass. Doing it once is most of a saved hour.
- Batching by step. Research four videos, then write four, then record four. Context-switching between modes is a real tax.
- Deciding faster, not editing faster. Most editing time is deliberation, not operations. A rule you follow (“cut every clip at the sentence”) beats a faster editor.
The one row a tool can actually move
The capture-and-assemble row is the exception to “tool-shopping is procrastination”, so it is worth being precise about what changes and what does not.
A script-to-video tool such as Flikifliki.ai · affiliate takes the finished script, renders the narration and cuts stock clips to each sentence in one pass. The 1–3 hours of assembly becomes minutes of machine time, plus however long you spend replacing the clips it chose. That is a real change to the table above.
It is not a change to the other seven hours. Research, writing for the ear and listening back are untouched, and the edit-and-tighten row shrinks only if you accept the first cut. Unedited stock clips matched sentence by sentence is also the interchangeable look the inauthentic content policy is aimed at, so the time you save on assembly is time you should expect to put back into making the video yours.
What it costs, read from Fliki’s pricing page on 2026-09-07: the first plan with commercial rights and no watermark is Standard, $28/month, or $21/month billed yearly, with 180 credits a month, which Fliki’s own FAQ puts at 60–90 minutes of finished video once re-renders are counted. The free plan carries a watermark, no commercial rights and a one-minute export cap — a trial, not a setup, for the same reason ElevenLabs’ free tier is. Set against the $6 starter setup, that is three to five times the monthly cost to buy back one row of the table.
The other row, and the other tool
Fliki collapses capture and assemble. It does not touch edit, caption and tighten, which is a separate 1–3 hours in the table and, for a narration-led video, usually the more stubborn one.
That row is what Descriptdescript.com · affiliate is aimed at, and the mechanism is genuinely different from “a faster editor”. Your video is a script read aloud, so the transcript and the timeline are the same object: delete a sentence in the text and the footage goes with it. The deliberation that eats most editing time — should this bit stay — happens in a medium you can read at a glance instead of scrubbing. It also hands back the transcript, so captions and the repurposed post come free rather than as another step.
Be precise about the limit, though, because it is the same limit as Fliki’s. It moves one row. Research, writing for the ear and listening back are the seven hours, and no editor touches them. And it is a speed purchase in the strictest sense: DaVinci Resolve edits the same video for nothing, so what you are buying is the hours, not the output.
What it costs, read from Descript’s pricing page on 2026-09-10: the first watermark-free tier is Hobbyist, $16/month billed annually or $24 month to month, exporting 1080p with 10 hours of media a month. The free plan watermarks the export and its AI credits are one-time rather than monthly, so it is a trial, not a setup. Against the $6 starter setup, that is 3.7x the monthly cost to buy back a different row of the table.
The honest way to decide between them: Fliki if assembly is your bottleneck, Descript if the edit is, neither if — as is usually the case — it is the research.
The honest bottom line
A faceless channel is cheap and it is not fast. Those are different claims and the marketing routinely merges them.
If the appeal was low cost, the numbers hold up completely — under a dollar a video, under about $50 a month all in. If the appeal was passive income, budget 5–12 hours per video before deciding whether the format is for you. Discovering that at video nine is the expensive way to learn it.
Related
- What a faceless channel actually costs per month — the money side of this page
- How long should a faceless video script be? — where the 1,200 words comes from
- What a minute of AI narration actually costs
- Do you need to pay for screen recording software?
Caveat: the times above are a reasoned model, not data from our own production. Treat them as a starting frame and replace each figure with your own after three or four videos — that measurement is worth more than any estimate here.
Sources
- Fliki pricing — plan prices, credits, export caps and the free plan’s watermark and commercial-rights lines