> For the complete documentation index, see [llms.txt](https://guide.hypercape.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://guide.hypercape.com/revelio/create-videos.md). # Create videos ## Choose the idea 1. Open a product from **My products**. Ideas are made on its **Create content** tab; **Analysis** and **My content** are the other tabs. The **Research basis** panel at the top of the tab shows where the product analysis comes from (see [Add a product](/revelio/add-a-product.md)). 2. Under **1 · Choose a template** (**What would you like to create?**), choose a template. Templates are grouped as **Show off the product**, **Explain and inform**, **Draw viewers into a story**, **Led by a character**, and **Graphics and card news**. Each card shows example scenes; a card for a story format also shows about how long the video runs, such as **About 25s**, and **A day in the animal protagonist's life** is marked **Needs an animal lead**. The examples only show a scene sequence: your video is planned new for your product and request, so results will vary. **Pause example motion** stops the moving previews, and **Play example motion** starts them again. 3. Choosing a template only sets the planner's defaults; it does not start a paid request. **Motion graphics** opens the motion graphics creator (see [Create motion graphics](/revelio/create-motion-graphics.md)), and **A swipeable story** opens the carousel planner (see [Create carousels](/revelio/create-carousels.md)). Once you choose, the gallery folds into **1 · Your template**; use **Change template** to pick another. Each template remembers the choices you made under it, while **Content you want** is shared across templates. 4. In **2 · Develop your idea** (**Add your own direction**), describe **Content you want (optional)**. Leaving it blank gives fresh AI ideas tailored to your product. If the purpose needs campaign details, fill in the required fields. 5. The story templates in **Draw viewers into a story** and **Led by a character** show **Choose a character** (or **Choose an animal lead** for **A day in the animal protagonist's life**) so you can select a saved lead, or **Create automatically** where it is offered. For the templates in **Show off the product** and **Explain and inform**, the picker appears once you turn on **Include an AI person** under **Customize**. The picker lists only the saved leads that the template's format can star. Animal characters and your own characters (see **Use your own character** below) use this same picker. **Talking product** and **Sensory ritual** are led by the product alone and cast no lead. 6. **A day in the animal protagonist's life** needs one of your saved animals and has no **Create automatically**: **Create ideas** stays unavailable until you choose one. If you have none, create one under **AI Influencer** first. 7. **Customize** opens more settings: **Content purpose** (**Product Introduction · Reviews**, **Ingredient emphasis**, or **Event Notice**; **More types** adds **Promotion · Discount**, **How to Use · Routine**, **Comparison · Selection Guide**, and **Brand Story**), **Story starting point**, **Include an AI person** (hidden for formats led by the product alone), **Campaign details**, **Production Method** (**Basic**, **Promote multiple products together**, or **Shooting guide (for real creators)**), and, where your account can choose models, **Generation model**. **Close customization** hides them again. Choosing **Customize** before any template selects **Product in the spotlight**. 8. Review the estimate and choose **Create ideas**. When ideas already exist, the button reads **Create fresh ideas**, and **View existing ideas** goes back to your list; from the list, **New concepts** opens the planner again. ## Choose a story format and language Every new idea is planned as a story: a lead who wants something, an obstacle, and a turn where the product pays off. When the template you chose is itself a story format, the planner shows that format; use **Change template** to switch. The other templates (**Product in the spotlight**, **Texture discovery moment**, **An ingredient story**, **Everyday routines**, **A character-led story**, and **A day in the animal protagonist's life**) show a **Story format** menu. Its empty choice, **AI pick (a fitting format for each idea)**, lets the AI pick a format for each idea to suit the product and current short-form trends. **Dialogue and caption language** sits next to it and applies either way. Choose both before creating ideas. | Format in the English UI | Length | Default voices and sound | | ----------------------------- | ------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Compact 4-Cut Story** | 4 cuts, about 20 seconds | Cuts 1–3: off-screen narration carries a short story arc (relatable hook, turning point, natural product moment). Cut 4: a silent visual payoff that can loop. | | **5-second social reel** | 1 cut, about 5 seconds | One silent cut that is the hook and the loop at once: it opens on a product moment and ends where it began, so it replays seamlessly. One short caption sits at the top of the frame; no dialogue or narration. The product is in frame, and a saved lead may hold it. | | **Interview twist** | 7 cuts, about 35 seconds | Cuts 1, 3, and 5: an off-screen interviewer asks questions while the lead listens. Cuts 2, 4, and 6: the lead answers. Cut 7: a silent visual payoff. | | **Short visual gag** | 2 cuts, about 10 seconds | Expectation versus reality, told through short on-screen text and action. Both cuts are silent: no dialogue or narration. | | **Character vlog** | 7 cuts, about 35 seconds | The lead speaks in the first person through cuts 1–6, followed by a silent reaction in cut 7. | | **Mini drama** | 7 cuts, about 35 seconds | Opens on a flash of the climax, builds a complication and a turning point, and ends on a next-episode hook. Mixes the lead's on-camera lines, off-screen narration, and one silent cut. | | **Character quest** | 7 cuts, about 35 seconds | The lead takes on an impossible problem and the product becomes the answer in a surreal scene; it ends on a loop. Mixes the lead's lines, narration, and silent cuts. | | **POV skit** | 5 cuts, about 25 seconds | "POV: you…" exaggerates a situation everyone knows, then the product solves it; it ends on a punchline. Narration and the lead's lines, with one silent cut. | | **Expectation vs reality** | 4 cuts, about 20 seconds | Sets the fantasy against reality and ends on a twist where the product bridges the two. Silent cuts with one narrated cut. | | **Transformation** | 6 cuts, about 30 seconds | Flashes the result first, then shows the routine and the change, and ends on a call to action. Narration with silent cuts. | | **Mystery reveal** | 6 cuts, about 30 seconds | Mysterious close-ups build curiosity and the product is revealed at the end. Narrated cuts between a silent opening and a silent loop. | | **Nature documentary parody** | 6 cuts, about 30 seconds | Films the lead like a rare animal in the wild, with off-screen narration in every cut, and ends on a punchline. | | **Sensory ritual** | 5 cuts, about 25 seconds | A product video told through texture, sound, and light, with no faces. Every cut is silent, with short on-screen text. | | **Myth vs truth** | 5 cuts, about 25 seconds | Exaggerates a common mistake, proves the truth with a metaphor scene, and ends on a call to action. Off-screen narration in every cut. | | **World inside the product** | 5 cuts, about 25 seconds | Dives into the product, travels through a world built from its key ingredient, and comes back with the result. Narration with silent cuts. | | **Talking product** | 5 cuts, about 25 seconds | The product itself gets eyes and a mouth and talks to the camera in every cut: a hello, the problem it solves, its key ingredient, how to use it, and a sign-off with a call to action. No person appears. | Each cut is approximately five seconds. Later edits, added clips, or transitions can change the exported duration. A silent cut has no speech or source audio. Background music is a separate choice made when you create the video (see **Generate and review**). **Short visual gag**, **5-second social reel**, and **Sensory ritual** are text-only: every cut is silent, captions are always on, and the confirmation shows **This format tells its story through captions and action, without dialogue.** Formats with fewer cuts use fewer clips; review the credit estimate before you create ideas or generate the video. **AI pick** never chooses **5-second social reel**. To make one, choose its card in **Show off the product** or pick it in **Story format** yourself. The list shows only the formats that suit the lead selected in **Choose a character**. With a saved person, it leaves out **Mystery reveal**, **Nature documentary parody**, **Sensory ritual**, **World inside the product**, and **Talking product**. With a saved animal or character, it leaves out **Sensory ritual** and **Talking product**; with **A character-led story** or **A day in the animal protagonist's life**, those two are left out even before you choose a lead. Otherwise, with no saved lead, every format is listed. **Sensory ritual** and **Talking product** are led by the product alone, so choosing one of them turns off and hides **Include an AI person**. Choose **Korean** or **English** in **Dialogue and caption language**. It controls the story's lines and on-screen captions, independently of the interface language. The language starts out as **English**, whatever the interface language; choose **Korean** to make the story in Korean. For **Short visual gag**, **5-second social reel**, and **Sensory ritual**, it controls the text only because there is no spoken dialogue. It does not translate every field of the posting package. Concepts created before story formats keep their saved structure (ordinary concepts: seven cuts, approximately 35 seconds); choosing a character alone does not convert an older concept into a story format. To change format or story language, return to planning and create ideas with those selections. Each format follows its own story beats, and the product appears only in the cuts the format marks for it. The formats do not use the ordinary product-introduction sequence or a mandatory sales outro; **Transformation** and **Myth vs truth** end on a call-to-action cut. For animal characters, do not depict human cosmetics being fed to or applied to an animal. ### Character story only, no product When a saved animal or character is selected in **Choose a character**, the planner also shows **Character story only, no product**. Turn it on to make an episode about your character with no promotion: the product does not appear on screen or in the lines, and the ideas are planned without the product's details. The planner summary then shows **No product** instead of the purpose and template. **World inside the product**, **Talking product**, and **5-second social reel** need the product, so they are not offered for a story without it, and **Character story only, no product** is hidden while one of them is selected. ### Episodes with no product at all A character or animal lead can also have its own episodes that belong to the lead rather than to a product. Where your account has access to **AI Influencer**: * Choose **Create video**, then **Create without a product** on the product list. Under **Who's the lead?**, pick a ready character or animal (or **New lead**). * Or open the lead's editor and choose **Make episodes** under **Episodes**. The front angle image must be filled first; only character and animal leads have their own episodes. On the episodes page, describe the story you want under **What story should we tell?** or leave it blank for ideas based on the lead's Soul and current trends. Choose a **Story format** (or leave it to the AI; the formats that need a product are not offered) and **Dialogue and caption language**, then **Create ideas**. Make a video from an idea as described below. ## Use your own character Besides an AI person or an animal, a story can star your own original character, such as a brand mascot, an imaginary creature, or a living object. Leads are made under **AI Influencer** in the left menu (on a phone, under **More**); menu items depend on account access. 1. Start a new one and set **Subject type** to **Character (mascot or creature)**. Choose a **Character type** and an **Art style** (**Photoreal** by default), and describe its **Look** in a sentence or two (up to 240 characters). The look is required. For a **Brand mascot**, you can add a **Brand name** (up to 40 characters): it appears as letters in the last scene of the mascot's episodes without a product; leave it empty to leave it out. 2. Optionally add **Your own character images (front, 3/4, side, back)** (for an animal: **Your own animal photos (front, 3/4, side, back)**): JPG, PNG, or WEBP, up to 8 MB each. Each slot has **Upload** (**Change** once it has an image) and **Cloud**, which imports an image from your Creator Cloud folder. With images added, the button reads **Create with uploaded images**; otherwise **Create** has AI draw the front image after you review its credit estimate. 3. **Character speaks on camera**, under **Adjust character traits (optional)**, is on by default, so the character says its lines on camera in dialogue scenes. Turn it off for a mascot with no mouth: the character never speaks, and its lines come as narration and captions. 4. A new character or animal opens in its editor right after it is created. Under **Angle images**, fill **Front**, **3/4 view**, **Side**, and **Back**: **Upload** or **Cloud** an image, choose **Make with AI** (or **Redo with AI**) to draw that angle (the front from the look you wrote, other angles from the lead's saved images), or **Clear** it. Images save as soon as you add them. Review the credit estimate shown before each AI image. 5. Under **Voice**, describe the voice in **What does the voice sound like?** (gender, age, and tone, up to 80 characters), then choose **Make the voice**. Revelio picks the best-fitting voice and plays a short sample. **Hear another voice** tries the next match, and **Remake with this description** uses an edited description. Each try is a small paid request that takes about 10 seconds (check the estimate shown), and it is saved at once, replacing the previous voice. Narration in every video of this lead uses this voice, read naturally rather than like a dub; on-camera lines and AI scenes get a closely matching voice. For a character that never speaks, every narration in its videos is read in this voice. AI people and animals have the same **Voice** section in their editors. 6. Write its **Soul (your lead's bible)**: personality, background, voice, story ideas, and rules to keep. **Insert outline**, shown while the Soul is empty, adds a starting structure, and **Load a Soul file (.md)** loads a Markdown or text file up to 256 KB. A Soul can be up to 20,000 characters. 7. Choose **Save** to store the text; saving uses no credits. Images and the voice are saved as you make them. Changes apply from your next video. To change them later, choose **Edit images & Soul** on the lead's card. Ideas and videos with that lead are planned with its Soul and angle images so it keeps the same design from every camera angle. Select it in **Choose a character** when you plan ideas. ## Review the scenario * Review **Recommended ideas**; expand **See other ideas** for the alternatives. * Open **Adjust scenario**, then **Write a scenario** or **Edit scenario**, depending on whether a draft exists. * Describe **The flow you want**. **Create scenes with AI** is a separate credit-using action; check its estimate. **Write my own** opens manual scene editing. Leaving the scenario empty lets AI build it when the video is made. * Use each scene's **Edit** control for its line. **Detailed editing (advanced)** contains scene directions and voice settings. A manually completed draft needs both a line and a scene direction for every cut. * For these story formats, keep each cut to at most **10 English words** or **24 Korean letters/digits**, excluding spaces and punctuation. In **Compact 4-Cut Story**, the narrated cuts (1–3) have a stricter limit of **8 English words** or **18 Korean letters/digits** so each line fits its five-second cut; the silent payoff cut keeps the standard limit. The one caption of **5-second social reel** allows **8 English words** or **16 Korean letters/digits**. The hint under each scene shows its limit, and a scene over the limit is marked in red until you shorten it. The newer formats (**Mini drama** through **Talking product**) allow **9 English words** or **20 Korean letters/digits** in a spoken cut (narration or the lead's line) and **8 words** or **16 letters/digits** in a silent on-screen headline. In **Speaker in this shot**, choose **Off-screen narration** (**Off-screen interviewer** in Interview twist), the lead's on-camera line (**Lead's line**, or **Animal protagonist** for an animal), or **Silent · On-screen captions** where the format permits. **Product appears in this scene** (**Show the closed product as a prop** for an animal) marks the product cuts; it is hidden in a character story without the product. When the planner named a direction for a cut, it is shown as **Direction**, for example its camera move, light, and transition. Short visual gag, 5-second social reel, and Sensory ritual stay silent; Compact 4-Cut Story ends on a silent payoff. * If a recorded line still runs longer than its cut, Revelio rewrites that line shorter and records it again instead of failing the video; the captions show the rewritten line. * Choose **Save** to keep your scenario, or **Cancel** to discard your changes. ![Manual scenario editor scrolled to show all seven scene rows, Detailed editing, Cancel, and Save](https://3720355098-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKmVnLoHDk5ZkU3shqZqa%2Fuploads%2F4h8oeNwV89rlT3VCuqL3%2Frevelio-responsive-2026-10-07-17-scenario.jpg?alt=media) ## Generate and review Choose **Make a video from this concept**. In the **Create video** confirmation, review the settings and the credit estimate (open **Details** to see what it includes), then choose **Create video**. For a story video, the confirmation also shows **Dialogue and caption language**, so you can switch between **English** and **Korean** before creating it. It is not offered when you continue a stopped video or when you wrote the scenario yourself; those keep their language. **Add background music** is on by default: it makes music that fits the video and plays it softly under the voices. Turn it off to make the video without music; the estimate updates to match. If the music cannot be made, the video is finished without it and the result says so; you can add music in the editor. The estimate covers each scene generation plus any required reference images, planning, and narration. Revelio edits in roughly five-second scenes; if a selected model only supports a longer clip duration, the estimate uses that actual requested duration. If the estimate is 1,000 credits or more, or a selected model has premium pricing, a red **High-cost generation warning** (or **Very high cost warning**) appears. Tick **I understand this is a high-cost generation** before **Create video** becomes available. For ordinary concepts, **AI person speaks on camera** switches between on-camera dialogue and narration, including for **Promote multiple products together**. It initially follows whether the concept has an AI person, then your last choice. Story formats use the saved per-cut speakers instead of this all-video switch; **Short visual gag**, **5-second social reel**, and **Sensory ritual** remain silent. Check subtitle settings when offered. Follow progress in **My content** before starting another request. Completion means that the generation and file-processing steps finished; it is not an aesthetic quality verdict. Review anatomy, movement, product details, and speech yourself. Revelio does not run paid visual-quality checks or automatically regenerate a completed model output to correct its appearance. A usable generated video remains a completed, billable result even if you are not satisfied with its quality. If you choose another version or remake a clip, review the new credit estimate before confirming that paid request. For higher quality, compare the advanced models available to your account and use a clear reference and a simple scene direction. A more expensive model can perform differently; no model guarantees a flawless result. Model selection is available only where your account's configured options permit it. If a video stops, its result in **My content** says what happened and what to do next. When finished scenes or prepared parts were saved, choose **Continue** to pick up from them; otherwise choose **Remake**. Provider errors, unavailable or unreadable files, and storage/processing errors remain technical failures. Read the error on the result in **My content** and check [Credits & help](/revelio/credits-and-help.md). Retrying a download or resuming a queued job reuses the existing job; it does not request a quality-correction video. Google Veo was already supported through Revelio's OpenRouter route. This update did not add it as a new integration. Model choices depend on account permission and the configured catalog; the model picker is shown only where your account can choose. No particular model guarantees the same result as Google Flow or eliminates all anatomy errors. **Next:** [Preview, edit, and download](/revelio/edit-and-download.md).