Faceless YouTube channel with AI video and voice — a full workflow

A faceless channel is a production line, not a performance. Nobody is on camera, so the three things that carry the video are the script, the narration and footage that does not look like the same stock library everyone else licenses. This is the workflow we run for that kind of channel, tool by tool, with the credit arithmetic at the end — computed from the pricing page, not from vibes. If you want the short version first, the faceless YouTube use case has it.

The shape of a faceless video

Before touching any tool, agree on the shape, because the shape decides the budget:

• Five to eight minutes of runtime, narration first, visuals second.
• Most of the runtime is stills with a slow pan added in your editor. Motion is expensive; use it where the story needs it.
• Eight to twelve AI clips of 5–10 seconds each as the "moments" — an establishing shot, a reveal, a transition.
• 16:9 everywhere. Set the format once in the video tool's Settings and never think about it again.

Do that and a video costs about what a lunch costs. Try to generate eight minutes of continuous motion and it costs what a weekend costs. The numbers are below.

Step 1 — the script, in Text

Open Text, pick RP+ (3 credits per completion) and the scene register. The scene register writes immersive prose, and immersive prose reads aloud far better than the bullet-point outlines a generic assistant hands you. Give it a beat sheet — the hook, three sections, the closer — and ask for one section per completion so you can steer between them.

Two habits that pay off downstream. Write for the ear: short sentences, one idea each. And punctuate on purpose, because the voice engine reads punctuation as delivery — a comma is a breath, an em dash is a pause, an ellipsis is a hesitation. A script drafted and tightened in four to six completions costs 12–18 credits. This is the cheapest part of the pipeline by a mile.

Step 2 — the narration, in Voice

In Voice, pick Voice One (3 credits per 100 characters, MP3 out). The roster has Aria (warm, intimate) and Orion (low, calm, and the one we default to for explainers), or you clone your own voice from a recording so the channel sounds like a person rather than a preset. Speed and expressiveness are sliders; the punctuation you wrote in step 1 does the rest.

Budget it by characters, not words. English prose runs about six characters per word including spaces, so a 900-word narration is roughly 5,400 characters. Voice One: 54 × 3 = 162 credits. If you are iterating on the script and only need a scratch track, Voice Turbo does the same voices at 2 credits per 100 characters, so the same narration is 54 × 2 = 108 credits. Generate the final read once, with Voice One, after the script is locked.

Step 3 — the footage, in Film mode

Open Film mode. It is a progressive storyboard: you add one scene at a time, each renders immediately, and the timeline underneath grows to a maximum of twelve. Every scene can be re-rendered, edited, reordered or duplicated, and each one can shoot up to four takes so you keep the best. The assembled player plays the whole thing in order, and drafts save as "My films".

Pick the engine deliberately. Motion One is on every account, including free, at 100 credits for 5 seconds and 180 for 10. Seedance 1.0 Lite unlocks at Hobby and costs 60 credits for 5 seconds and 110 for 10. Clip length is capped by your plan — 10 seconds free, 15 on Hobby, 20 on Creator, 30 on Studio — and also by the engine, since Motion One renders 5 or 10 seconds only.

Write each scene as one flowing paragraph: subject, motion, camera, light, mood. This is a scene we would use in a maritime-history video:

A lone lighthouse keeper climbs a spiral iron staircase inside a stone tower at dusk, lantern in hand, slow push-in following him upward, warm lantern glow against cold blue light from the slit windows, dust hanging in the air, calm tempo, documentary feel, 35mm film grain, wide 16:9 frame.

The director's rack gives the same result with fewer words: Film type "Documentary", Tempo "Calm", Camera gear "35mm film", Palette "Warm film", Lighting "Practicals". Set it once and it folds into every scene's prompt. When a scene comes out flat, hit Enhance — it rewrites your rough draft into a director-grade prompt and costs nothing.

Step 4 — the stills, with Image One

Most of the runtime comes from stills. In Image, use Image One (10 credits each) at 16:9, batch four at a time, and put "text, watermark, logo" in the negative prompt so you do not spend an evening cloning out captions. Twenty stills is 200 credits. Keep the palette consistent with the footage — same "Warm film" wording in the prompt — and the cuts between stills and motion stop being visible.

The credit budget per video

Two recipes, both on Motion One unless noted.

Lean video — script 4 completions × 3 = 12; narration 162; stills 20 × 10 = 200; clips 6 × 5 seconds × 100 = 600. Total 974 credits, which fits inside one Starter pack ($10 for 1,000 credits, so roughly $10 per video at that rate).

Full video — script 12; narration 162; stills 200; clips 12 × 10 seconds × 180 = 2,160; a second take on four key scenes 4 × 180 = 720. Total 3,254 credits. Swap the clips to Seedance 1.0 Lite and the same video is 12 + 162 + 200 + 12 × 110 + 4 × 110 = 2,134 credits.

On a plan, that means: Creator ($49 for 6,000 credits) is six lean videos a month or almost two full ones on Motion One (6,000 ÷ 3,254 ≈ 1.8); Studio ($149 for 20,000) is about six full videos on Motion One (20,000 ÷ 3,254 ≈ 6.1) or nine on Seedance Lite (20,000 ÷ 2,134 ≈ 9.4). Failed renders refund themselves, and unused credits roll over, so a slow month is not a wasted month.

Assembly and publishing

Every render lands in your Library automatically, with its full recipe, so re-rendering the one scene that bothers you next week is a Remix, not a rewrite. Download the MP3 and the clips, lay the narration down first, cut the stills and clips to it, and export. If you cut in Premiere Pro or After Effects, the eroq panel generates from inside the app and imports renders straight into the open project. The film endpoints in the API docs let you script the whole storyboard render once you have a format that works.

FAQ

Can I test the whole workflow on the free credits?

Mostly. Fifty free credits cover a script draft (3), a 1,000-character narration test (30) and one still (10) — 43 credits. A single Motion One clip is 100 credits, so the video step needs a pack or a plan.

Which engine should a faceless channel default to?

Motion One if you are on the free tier or want the same engine for everything, including mature fiction channels. Seedance 1.0 Lite from Hobby upward when the budget matters more than the top resolution; it is roughly 40 percent cheaper per 10-second clip (110 versus 180 credits).

How long can one clip be?

It depends on the engine and your plan. Motion One renders 5 or 10 seconds; Seedance 2.5 goes from 4 to 30 seconds on Creator and up. Your plan caps the length regardless — free 10 s, Hobby 15 s, Creator 20 s, Studio and above 30 s.

Ready to shoot the first one? Open Film mode — the script and the voice are the cheap part, so start there.