In my latest Faceless Journey video I started a brand-new sleep channel from scratch, Bible mysteries and lost texts, for sleep, and generated an 80-minute video for it. A lot of you asked for the exact settings. Here they are, copy-paste ready.
They're built from the presets of my main channel, Sleepy Time Historian (175K subs), so they start from settings that already work. Swap the topic words for your own niche and keep the structure.
Everything below is plain text, so it works as a starting point in any tool. In AutoTube each block goes into the matching preset field.
1. Channel brief
The channel brief is what ideation and every later step read. Keep it to one paragraph: what the channel is, and what to avoid.
Bible mysteries and lost texts, told slowly for sleep. Long, calm videos of about 90 minutes on the ancient Near East and early Jewish and Christian history: lost and excluded books (Enoch, Jubilees, the Gospel of Thomas), the Dead Sea Scrolls, the Nag Hammadi codices, the 400 years between the testaments, vanished peoples and cities of the Bible, and how the Bible's texts were copied, translated and gathered. Every video is built on real surviving manuscripts, inscriptions and artefacts, presents scholarly views as views and names where traditions disagree. Respectful to believers and non-believers alike; never sensational, never creepy. Titles promise a clear mystery or question, not shock. Avoid prehistory and Sumer (another channel's lane) and avoid repeating a topic already covered.
2. Script preset
Settings
| Setting | Value |
|---|---|
| Word count | 12,000–13,000 (about 80–95 minutes of narration) |
| Research | On (always, for long sleep scripts) |
| Outline model | GPT-6.1 Sol |
| Script model | Claude Sonnet 5.5 |
| Dual model | On: the first 2 sections on the script model, the rest on GPT-6.1 Sol to save cost |
| Outline | Automatic |
Why research on: a script this long without research drifts into filler. With it, the script is built on real sources, and those sources go straight into the video description.
Script requirements
GLOBAL:
- Tell this as one calm, engaging story that holds the listener's curiosity. Aim for accuracy and narrate with quiet confidence. Never invent facts, quotes, scholars, manuscripts, dates or numbers: use only what the research supports, and leave a point out if the sources are thin.
- Be respectful of every tradition and every listener: believers, doubters and the curious. Never mock, never debunk for sport, and never tell the listener what to believe.
- Present scholarly views as views. Say "according to the text", "one view holds", "another reading is", and name who holds each view when the research does (a tradition, a school, a scholar). Give dates as ranges when dating is debated. Where traditions genuinely disagree (for example the Jewish, Catholic, Orthodox, Protestant and Ethiopian Orthodox canons, or religious and secular scholarship), say so plainly in a sentence or two, describe each position fairly, and move on. Do not keep reminding the listener of what cannot be known.
- No sensational framing: no "forbidden", "banned", "suppressed", "the truth they hid", "cover-up" or conspiracy claims, unless a source documents that exact event, and then report it plainly.
- Bedtime tone: warm, unhurried and steady. No shocks, no cliffhangers that raise the pulse, no graphic violence. The energy relaxes gently as the video goes on: later sections are slower, more descriptive and more soothing. Avoid repetition or restating the same ideas; each section adds something new.
- No plugs of any kind: no products, shows, books, courses, websites, channels, apps, Spotify, links, offers or discounts. Do not ask for likes, subscriptions, follows, reviews or support.
INTRO/HOOK:
- Open with a quiet, curious hook that makes the listener want to stay.
- Within the first minutes, ask the listener ONCE, kindly and briefly, to comment where they are listening from and what time it is there, because it is lovely to feel connected across the world. This is the only request to the audience in the entire script.
CONCLUSION:
- Close softly: return to the image the story began with, leave the listener with quiet wonder, and wish them a restful night. No summary, no call to action, and no second request for comments.
Style guide (the tone of your channel)
VOICE
- One calm, warm narrator speaking quietly to one listener late at night. Plain spoken English with contractions ("it's", "we'll", "that's"). Never academic, never preachy, never dramatic.
- Reverent curiosity. The narrator respects the texts and the people who treasured them, and equally respects the listener who is a believer, a doubter or just curious. Wonder over drama: "isn't it remarkable that this survived?", never shock.
- Address the listener as "you" and fold them in with "we" and "us". Steady, low energy even when the subject is large.
OPENING
- Start with a quiet, concrete scene (a place, a time of day, a person with a task), then widen out to why it matters. Land the central fact once, simply.
- Name when and how we came to know it ("It would be centuries before anyone read it again"), so the video gently promises a reveal.
- Give the listener-comment prompt once, in the first minutes, as a warm aside in the narrator's own words with a small picture ("Somewhere it's evening, somewhere it's just before dawn"), then return to the story. Do not ask for anything else, anywhere.
SENTENCE RHYTHM
- Mostly medium, unhurried sentences, with an occasional very short one as a pause: "That's all." "A quiet hillside." Fragments sparingly.
- Soft spoken pivots: "Now,", "So,", "Let's go back.", "For now,". Vary them. No exclamation marks, no "You won't believe", no cliffhanger shocks.
- Lists of three for a gentle rhythm: "a hole in the earth, a jar, a handful of books."
VIEWS, TRADITIONS AND SOURCES
- Attribute everything contested. "According to the text...", "One view holds...", "Another reading is...", "Jewish tradition says..., while many Christian traditions say...". Name the tradition or scholar when the research does.
- Dates as ranges when debated ("between the third century BCE and the first century CE"). Say what is unknown once, briefly, then move on.
- Where traditions disagree, state each position fairly and without ranking them. Never mock a belief and never present a theory as settled. Never invent a quote, a scholar or a manuscript.
- Avoid "forbidden", "banned", "hidden from you", "what the church doesn't want you to know".
EXPLAINING
- One technical term at a time, named and explained in everyday words (codex, papyrus, canon, Septuagint, Coptic). Ground numbers and ages with a familiar comparison.
- Introduce real people and places by name and role. Tell a discovery as a small story: what they expected, what they found, what they did next.
IMMERSION
- Regularly step into second-person scenes: "Picture a scribe in a lamplit room...". Use ordinary sensory detail (clay, reed pens, lamp oil, cool stone, the smell of ink) and daily life. Then step back out to what the evidence actually shows.
STRUCTURE
- One continuous, flowing story, not a list of facts. Each section raises a question the next one answers, and ends on a gentle turn, never a jolt.
- As the video goes on, let sentences lengthen, imagery soften and pacing slow. Close by returning to the opening image and wishing the listener a restful night. No summary, no call to action.
AVOID
- Plugs of any kind, links, offers, and requests to like, subscribe or follow.
- Headings, bullet points, stage directions, anything that is not spoken narration.
- Crutches: "Imagine..." at the start of every scene, "fascinating", "delve", "tapestry", "testament to".
- Repeating the same fact or phrase across sections.
- Copying the sample's topic, facts or sentences. Match its voice and rhythm, not its content.
Style sample (about 400 words)
The style guide works best next to a short sample in the voice you want. Pick a topic you'll never make a video about, so the script copies the voice and not the content.
Somewhere in Egypt tonight, the Nile is moving slowly past a long line of limestone cliffs, just as it did on a December morning in 1945. A man is walking toward those cliffs with a mattock over his shoulder. His name is Muhammad Ali al-Samman, and he isn't looking for treasure. He's looking for soft, dark soil to spread on his fields. That's all.
Before we go any further, if you're listening to this to fall asleep, you're in exactly the right place. Take a slow breath. And if you'd like to, tell me in the comments where you're listening from, and what time it is there. Somewhere it's evening, somewhere it's just before dawn, and I like to think of all of us sharing one quiet story.
Now, back to the cliff. Near a large boulder, his tool strikes something hard. Buried in the earth is a sealed clay jar, about a metre tall. He hesitates. In the villages, people say a jar like this might hold a jinn, a spirit. But gold is also possible, and gold would change a great deal. So he lifts the mattock and breaks it open.
There's no gold. Inside are books. Thirteen of them, bound in leather, their pages made of papyrus.
At least, that's how he told it. And it's worth saying gently that this is his own account, given to scholars many years afterwards, and the details shift from one telling to the next. What nobody disputes is where the books ended up: in a museum in Cairo, where they're still studied today.
They're called codices, which simply means early books with pages instead of scrolls. Together they hold around fifty texts, all written in Coptic, the language of Christian Egypt. Many were translated from older Greek originals, and several were works that no one alive had ever read.
One of them is a collection of sayings attributed to Jesus, known as the Gospel of Thomas. Scholars still disagree about when it was first composed. Some place it in the first century. Others place it well into the second. Both views rest on careful arguments, and we'll hear them both.
For now, let's stay with the jar. A quiet hillside. A hole in the earth. The sun climbing slowly over the river. And, folded inside the dark, a library that someone, long ago, chose to bury.
3. Voice preset
I designed the voice from a plain-words description (Gemini voice design, Male, English UK), listened to a few takes and kept the best. It's one of the cheapest voices available, and 80 minutes of narration comes back in a few minutes.
A British male documentary narrator in his forties with a warm, rich baritone and a clear received-pronunciation accent. He speaks at a natural documentary pace, measured and confident, never slow or drowsy. Clear diction, gentle authority and quiet curiosity, like the narrator of a BBC history documentary. Steady and even, with light natural emphasis on names, places and dates, and no theatrics.
Speed 1.00. Tip: never put "sleep", "bedtime" or "slow" in a voice prompt. The model overacts them and the voice comes out far too slow. The calm comes from the script, not the voice.
4. Visuals preset
Mix: Advanced mode, split into segments. The opening (~80 s) is all AI images. After that, roughly half AI images, half stock. When a line names a real manuscript or artefact, the stock slot pulls a real image of it from Wikimedia Commons. No AI video: that's where costs explode, and sleep videos don't need it. Image models: Lucid Origin for the opening, Ideogram for the rest.
Visual guidance
Painted illustration in the manner of a dark, softly lit oil and watercolor painting: muted earthy ochre, umber, deep indigo and dusty blue, warm candlelight and oil-lamp glows against deep shadow, visible brush texture, low detail, gentle vignette. Calm and dark, never bright or harsh.
The content we make is generally about the ancient Near East and early Jewish and Christian history: Egypt, Mesopotamia, Canaan, Israel and Judea, the Roman-era Mediterranean, desert caves and monasteries, scribes, scrolls, codices, clay jars and temples.
IMPORTANT: Lead with WIDE, expansive landscapes and wide perspectives: desert cliffs, the Nile, hillsides, ruins at dusk, lamplit interiors seen from a distance. AVOID close-ups of faces and individual people; keep human figures small, distant, from behind, in silhouette or reduced to hands and objects. Never depict God or Jesus. No text, lettering or symbols in the images. Yet still try to make it voiceover-relevant - be creative.
Motion graphics style
We make Bible mysteries and lost-texts documentaries to fall asleep to. Go for a darker, candle-lit look instead of bright and blinding.
5. Render settings
| Setting | Value |
|---|---|
| Resolution | 1280×720, 24 fps |
| Subtitles | On: Roboto italic, gold (#dcb218), word-fade highlight |
| Motion | Slow zoom 1.2, random pans and zooms |
| Transitions | Fade, push, wipe, gradient wipe, dip to black |
| Reveal on long scenes | On for scenes of 4 s or longer, on half of them |
6. Description preset
Description requirements
This is a calm YouTube channel about Bible mysteries and lost texts, made to fall asleep to. Start with a gentle, thought-provoking question about the topic. Add a few lines of basic background that build curiosity without exaggeration.
Keep it short and to the point, in short paragraphs. Use SEO-optimized keywords naturally here and there. Appeal to a broad, respectful audience of believers, doubters and the curious. Describe scholarly views as views; never claim anything is hidden, forbidden, suppressed or proven. No links, no requests to buy, follow or subscribe, and NO emojis.
Tags requirements
Generate tags related to the Bible mysteries and lost texts for sleep niche, e.g. "bible stories for sleep, lost books of the bible, biblical history for sleep, sleep documentary"
Description template
{generatedDescription}
Chapters:
{chapters}
Sources:
{sources}
Credits:
{attribution}
Note on accuracy:
This video is built from the research sources listed above. It is narrated with an AI-designed voice and illustrated with AI-painted images together with the real images and footage credited above. We aim to be accurate and respectful: scholarly views are presented as views, and where traditions disagree we say so. This is storytelling for relaxation, not a formal academic or religious source, so please check the sources yourself.
Tip: with research on and lots of Wikimedia images, the credits block can push the description past YouTube's 5,000-character limit. Trim the credits to the essentials (author, licence, link) before you publish.
What it cost
The whole 80-minute video came to about 670 credits: script with research 139, voiceover 268, images and stock 214, prompts and director 45, and the render was included in the daily allowance. Depending on how you buy credits, that's roughly $3–6.
Watch the full walkthrough, including how I picked the niche: How I got my sleep channel remonetized, plus starting a new one from scratch



