Turn text or an approved image into a close-up ASMR clip with directed motion and matching sound. Review each result before publishing.
1. Open a Model and Select Text to Video
This workflow was checked in the signed-in APOB AI Create workspace on August 31, 2026. Controls, quality tiers, and credit estimates can change, so use the labels shown in your account as the final source of truth.
Your goal: Reach the generator with a character context available if the concept later needs one.
Common problem: Opening the general creation URL without first selecting a model can leave the workspace waiting for content context. Starting in the wrong mode can also lead users to search for sound controls that belong to a different workflow.
In APOB AI: Sign in at app.apob.ai, open a Personal or available Community model from Home, then select Video → Text to video in the left navigation. The Create workstation guide explains that the mode menu controls which inputs and settings appear.
Realistic example: A faceless Shorts creator opens a model, switches from Image to Text to video, and prepares a five-second crystal-fruit test without uploading a source frame.
Quality tip: Decide whether you need a recurring character before prompting. If the subject is only a glass pear or jelly keyboard, text-to-video keeps the test simpler.
Expected output: The Text to Video composer is visible, with the prompt area, Native Audio control, presets, output settings, and Generate button.
2. Enable Native Audio and Write an Action–Material–Sound Prompt
Your goal: Describe a visible event and the exact sound it should produce.
Common problem: “Make a relaxing ASMR video” leaves the object, contact point, camera, and sound undefined. Several actions in one short prompt can produce unrelated audio or morphing objects.
In APOB AI: Keep Native audio on, then write the prompt as subject → material → action → camera → lighting → sound timing → exclusions. According to the current APOB Text to Video guide, Native Audio is available across Text to Video quality levels and is required in Ultra S.
Realistic example — copy-ready prompt: “Macro close-up of a chilled pear made from pale green translucent glass on a dark slate board. A polished chef's knife completes one slow straight slice. Locked camera, cool side light, realistic refraction. The sound begins at visible contact: one crisp crystalline crack, a short scrape against slate, then silence. No voice, no music, no text, no logo, no extra hands, no broken knife, no camera shake, no object morphing.”
Quality tip: Give the movement a visible start and finish, and tie the sound to contact. A single slice plus a single crack is easier to evaluate than cutting, pouring, whispering, and moving the camera in the same five seconds.
Expected output: A prompt in which the primary motion and primary sound describe the same event.
3. Set the Frame, Duration, Resolution, and Quality
Your goal: Match the test to its publishing channel and understand the cost before generation.
Common problem: A wide frame can crop poorly on TikTok, Reels, or Shorts, while an expensive first run may be wasteful before the prompt is validated.
In APOB AI: Choose the aspect ratio, duration, resolution, and quality shown below the composer. In the live test, 9:16, 5 seconds, 720P, Ultra S, and Native Audio displayed 200 credits per second and 1,000 credits on the Generate button. This is a dated observation, not a fixed price; the live button is authoritative.
Realistic example: The glass-pear test uses 9:16 so the knife and fruit remain inside a mobile safe frame. A widescreen rain-window insert would use 16:9 instead.
Quality tip: Validate composition and sound direction with the shortest suitable test before extending the concept. Check the cost again after changing quality, duration, resolution, or audio options.
Expected output: A clearly framed test with a visible, accepted credit estimate and no accidental crop of the contact action.
4. Generate and Review the Result in the Feed
Your goal: Decide whether the clip passes both visual and audio review.
Common problem: A thumbnail may look correct while the full clip contains an early impact, an unwanted voice, changing labels, deformed hands, or an incomplete action.
In APOB AI: Select Generate only after confirming the displayed cost. The job appears in the workstation feed, and APOB notifications report completion or failure. Open the result, unmute it, and inspect action completion, object/tool geometry, sound identity, and contact timing. The APOB troubleshooting guide says credits for a failed generation are returned automatically; a technically completed but unusable variation still requires a new creative decision.
Realistic example: If the pear splits correctly but the crack occurs before the blade touches glass, revise the sound line to “sound begins exactly at visible blade contact,” remove unnecessary effects, and run a new variation.
Quality tip: Use headphones and review at normal speed and frame by frame. For a loop, trim and align the approved clip in an editor rather than assuming viewer replay is seamless.
Expected output: One clip that passes a documented check for subject continuity, geometry, action, sound timing, unwanted speech/music, disclosure, and crop safety.
5. Switch to Image to Video When Identity or Composition Matters
Your goal: Animate an approved product, fictional host, or prepared scene instead of asking text-to-video to invent the first frame.
Common problem: Text-only runs can change a bottle, face, label, or palette between attempts. A weak source image can also carry visual defects into motion.
In APOB AI: Select Video → Image to video, then choose existing content or upload a reference image. Describe only how that image should move. Supported setups can add a last frame, Native Audio, an Element, a Voice model, or camera direction. The official Image to Video guide notes that a last frame and multi-shot cannot be used together; the source image establishes the initial composition.
Realistic example: A skincare team uploads an approved 9:16 serum-bottle frame, requests one slow pump and a soft mechanical click, and keeps labels under human review. A fictional tea host can instead use a consistent authorized model and one restrained hand action.
Quality tip: Correct distorted fingers, unreadable packaging, or impossible reflections before animation. Use a Voice model only for an original or authorized voice and only when speech is necessary.
Expected output: A motion test anchored to a recognizable first frame, with the selected audio and reference controls visible before generation.
Ready to test one trigger? Create an AI ASMR video with one subject, one material, one visible action, and one matching sound cue. Review the live credit estimate before generating.
Expert Prompts for AI ASMR
Use one prompt card at a time. Reference-led prompts assume you have already selected the matching product, character, or source image.
1. Surreal glass-fruit cutting
User: Faceless short-form creator Purpose: Test a recognizable impossible-material trigger. Copy-ready prompt: “Macro close-up of a chilled pear made from pale green translucent glass on a dark slate board. One gloved hand holds the pear steady while a polished chef's knife completes one slow, straight slice. Locked camera, shallow depth of field, cool side light, realistic refraction. The sound begins at contact: one crisp crystalline crack, a short clean scrape against slate, then silence. No voice, no music, no text, no logo, no extra fingers, no broken knife, no camera shake, no object morphing.” Expected output: One complete slice with a clear impact and stable object. Channel: 9:16 Shorts/Reels/TikTok. Customize: Replace pear/glass/slate with one new object–material–surface combination per test.
2. Skincare pump and serum texture
User: Ecommerce skincare marketer Purpose: Turn an approved product image into a tactile detail clip. Copy-ready prompt: “Use the supplied serum bottle as the exact first-frame product reference. Macro product shot. A clean fingertip presses the pump once; one clear droplet lands on a glass dish and slowly spreads. Preserve the bottle shape, cap, label placement, and brand colors. Soft diffused bathroom light, locked camera. One quiet pump click and one soft liquid tap, low room tone, no voice, no music. No label mutation, no extra hand, no duplicate bottle, no medical claim, no camera movement.” Expected output: A controlled pump-and-drop moment that keeps the packaging recognizable. Channel: 9:16 Reel or product-page microvideo. Customize: Match the droplet viscosity to the real product; do not imply a texture the product does not have.
3. Fictional tea host with one whispered line
User: AI influencer creator Purpose: Build a recurring, disclosed character-led ASMR series. Copy-ready prompt: “Use the selected fictional adult host as the same tea presenter in a green reading room. Medium close-up, three-quarter view. She places a ceramic lid on a small teapot, pauses, then looks toward camera. Soft warm lamplight, minimal head movement, stable facial identity. Ceramic contact and a light fingertip tap. In a slow natural whisper, she says once: ‘Let the first cup settle.’ No music, no extra dialogue, no subtitles, no exaggerated smile, no extra fingers, no camera shake.” Expected output: A recognizable host, one tactile action, and one short authorized/designed voice line. Channel: 9:16 episodic Reel or Short. Customize: Keep the set and host constant; change only the weekly object and action.
4. Satin-fold fashion detail
User: Fashion brand or designer Purpose: Show material movement without a full model shoot. Copy-ready prompt: “Macro overhead view of an ivory satin scarf from the supplied product image. Two clean hands complete one slow fold from left to right, preserving the printed border and fabric color. Soft window light reveals the weave and highlight roll-off. Locked camera. Gentle cloth glide and one quiet fold, no jewelry noise, no voice, no music. No warped hands, no changing print, no new seams, no liquid texture, no fast motion.” Expected output: One controlled fold with believable fabric sound and preserved design. Channel: 9:16 fashion Reel or 16:9 campaign insert. Customize: Replace satin with the real fabric and adjust the sound description accordingly.
5. Cold-brew can opening
User: Beverage marketer Purpose: Prototype a high-salience product sound cue. Copy-ready prompt: “Use the supplied cold-brew can as the exact product reference. Extreme close-up of condensation on the lid. One thumb lifts the pull tab; the seal opens once and a fine mist escapes. Preserve logo, typography, can proportions, and metallic finish. Locked camera, bright rim light, dark café background. One metallic click, one short pressurized hiss, then quiet fizz. No hand distortion, no exploding liquid, no changing label, no voice, no music.” Expected output: A single clean open with synchronized click/hiss/fizz. Channel: 9:16 social ad test. Customize: Use the real packaging reference and ensure the generated behavior does not misstate the product.
6. Jelly keyboard press
User: Satisfying-video creator Purpose: Turn a related-search concept into an original tactile scene. Copy-ready prompt: “Top-down macro shot of a compact keyboard whose translucent keys are made from firm blueberry jelly. One fingertip presses a single center key until it compresses and rebounds once. Clean pastel desk, soft studio light, locked camera. A muted jelly press aligned exactly with visible compression, followed by one light key click. No typing sequence, no extra fingers, no melting keys, no text, no voice, no music, no camera movement.” Expected output: One readable compression/rebound cycle suitable for a loop edit. Channel: 9:16 TikTok or Short. Customize: Change key material and sound as a pair; do not keep the same audio description for foam, glass, metal, and jelly.
7. Rain-on-window ambient insert
User: YouTube ambience editor Purpose: Generate a restrained visual insert for a longer human-edited video. Copy-ready prompt: “Locked 16:9 close view from inside a quiet cabin at dusk. Rain droplets gather and travel slowly down a dark wood-framed window; a softly blurred pine forest sits outside. Warm lamp reflection in one corner, no people, no lightning. Natural light rain against glass with a distant low wind, no thunder, no voice, no melody, no sudden volume change, no camera movement, no changing window geometry.” Expected output: A stable short ambient shot with controlled room-scale sound. Channel: 16:9 YouTube insert. Customize: Generate several distinct approved inserts, then assemble them with licensed/original long-form audio.
8. Tissue-and-foil product unboxing
User: Small ecommerce seller Purpose: Create a packaging-focused ASMR hook before a campaign shoot. Copy-ready prompt: “Macro overhead shot of the supplied gift box on a clean cream table. Two hands lift one tissue fold and peel back one gold foil seal, revealing the approved product without changing its packaging. Soft daylight, locked camera. Gentle tissue rustle followed by one crisp foil peel; no music, no speech, no extra products, no unreadable label changes, no extra fingers, no fast unboxing montage.” Expected output: A two-beat packaging reveal with distinct paper and foil sounds. Channel: 9:16 Reel, TikTok, or product-launch Story. Customize: Match every package layer to what customers actually receive.
Generate the action and its sound in the same workflow
The central production problem in synthetic ASMR is not merely adding an audio track; it is making the sound correspond to the visible material and contact point. Native Audio lets the prompt describe both parts of the event. Availability depends on the selected workflow and quality, so the current toggle and Generate estimate remain the source of truth.
Start from a controlled product or character image
Text-to-video is useful for surreal one-offs, but a brand bottle or recurring host usually needs a visual anchor. Image to Video lets a team begin from an approved frame, reducing the risk that every test reinvents the subject before the action even starts.
Build a series around a reusable fictional host
A saved Portrait Model can support a recognizable fictional ASMR character across images and videos. That is more practical for a weekly tea, skincare, desk, or bedtime series than prompting a new anonymous face for each post. Consistency is still a review goal, not a guarantee—especially when hands, profiles, or fast motion dominate the shot.
Reference products and scene elements where the mode supports them
Elements are supported in Image to Video at Ultra and Ultra S and in Text to Video at Ultra S. A marketer can reference the intended bottle, pouch, mascot, or set instead of packing every visual detail into prose. If a quality switch removes unsupported inputs, APOB warns before clearing the current prompt and Elements.
See the cost before generating
APOB prices video by output second, with the displayed rate affected by plan, duration, resolution, quality, and options such as Native Audio. The Generate button shows the current cost before the run. This makes it easier to compare a short concept test with a higher-quality final attempt without relying on a fixed marketing-page estimate.
Continue into related video and voice workflows
An approved image can move to Image to Video or Talking Avatar, and an existing face video can move to Lip Sync. These are useful handoffs for a creator-led intro or a short whispered line, but nonverbal trigger content should stay in the simpler native-audio video workflow when speech is unnecessary.
Obtain consent for every real face and voice
Do not animate, clone, or lip-sync a real person without explicit, verifiable permission for the exact project and channels. The APOB Terms of Service prohibit appropriating a person's name, likeness, voice, or persona without consent and prohibit deceptive impersonation. A synthetic label does not cure unauthorized use.
Confirm input rights and commercial scope
Use product images, logos, fonts, scripts, photographs, music, recordings, and sound effects that you own or are licensed to process and publish. As between the user and APOB, the current terms assign the user rights in generated content to the fullest extent permitted by law, subject to third-party rights and the licenses granted to APOB. That is not a guarantee that every output is copyrightable or free of third-party claims.
Treat generated sound and uploaded sound differently
Native Audio is generated with the video. Uploaded music, ambient recordings, samples, or voice clips still need their own licenses and releases. Do not copy a recognizable sound recording or clone a voice merely because a prompt can describe it.
Disclose realistic synthetic media
TikTok requires labels for AI-generated content containing realistic images, audio, or video and prohibits certain harmful impersonations even when labelled. YouTube requires disclosure for realistic content that is generated or meaningfully altered, including making a person appear to do or say something they did not. Meta requires its disclosure tool for certain photorealistic video or realistic-sounding audio that was digitally created or altered. Use each platform's current upload controls and add a clear caption when context could still be misunderstood.
Keep advertising claims truthful
A generated serum spread, food texture, fabric fold, unboxing, or beverage reaction must not imply a product property that the real item does not have. Synthetic product endorsements also require the same sponsorship and material-connection disclosures as other advertising. For US campaigns, the FTC's endorsement guidance explains that unexpected material connections should be disclosed clearly; check the rules in every market you target.
Avoid sexualized or sensitive roleplay and protect minors
APOB prohibits adult/NSFW content. Do not use real or synthetic minors in sexualized, intimate, manipulative, or age-inappropriate ASMR scenarios. TikTok also prohibits certain realistic AI depictions of minors and the unauthorized likeness of adult private figures.
Review visibility and data terms before uploading confidential assets
Free-plan visibility and paid private mode are not the same as confidential processing. The APOB terms grant the company a broad service license over user inputs and generated content and explain that inputs may be shared with third-party providers needed to operate the service. Review the current Privacy Policy and terms before uploading unreleased products, client secrets, personal recordings, or biometric-like likeness assets.
Keep a human approval step
Review every frame and audio cue for safety, accuracy, unwanted speech, logos, product claims, and misleading identity. AI ASMR is a production tool, not evidence that an event occurred or that a product, person, or health benefit is real.
This section provides general production guidance, not legal advice. Review the current APOB terms and obtain qualified legal advice for high-risk likeness, voice, advertising, or cross-border commercial uses.
Surreal cutting loops for a faceless Shorts series
Who it is for: A faceless entertainment creator testing impossible materials. Task: Publish a sequence such as crystal fruit, cloud bread, glowing mineral candy, or layered paper planets. Production problem: Physical fabrication is expensive, while broad prompts produce random tools, cuts, and sounds. APOB workflow: Use Text to Video, define one object/material/action, keep the camera locked, and enable Native Audio. Approve each clip before editing several together. Format: 9:16 TikTok, Reels, or YouTube Shorts. Limitation: Repetitive, mass-produced uploads may feel inauthentic and can create monetization risk. YouTube's channel monetization policy says repetitive or mass-produced “inauthentic content” is ineligible, so add original selection, editing, sequencing, or commentary rather than publishing near-identical batches.
Product ASMR for ecommerce detail clips
Who it is for: Skincare, beverage, stationery, packaging, and lifestyle sellers. Task: Show a serum pump, magnetic cap, foil seal, textured label, condensation, fabric pouch, or unboxing layer. Production problem: A full reshoot for every hook, crop, or seasonal surface is slow, and a text-only tool may change the product. APOB workflow: Begin with an approved product image, animate one interaction, mention the expected sound, and use a product Element where supported. Keep labels and claims under human review. Format: 9:16 paid/organic social test or a short ecommerce detail clip. Limitation: The generated clip must not misrepresent product size, texture, function, included items, or performance.
A recurring fictional tea or bedtime host
Who it is for: An AI influencer creator building a recognizable series without using an unconsenting real person. Task: Publish quiet tea rituals, desk resets, book handling, fabric brushing, or soft-spoken fictional roleplay. Production problem: One-off faces weaken series recognition, but aggressive character motion can introduce identity and hand drift. APOB workflow: Create an original adult Portrait Model, generate a clean first frame, animate one controlled action, and add an authorized or designed voice only if speech is essential. Format: Reels, TikTok, or Shorts episode. Limitation: Some ASMR audiences value visible human effort and connection. Label the synthetic host and build a clear creative premise instead of pretending the character is a real creator.
Fabric and accessory ASMR for a fashion lookbook
Who it is for: Fashion designers, accessory brands, and creative agencies. Task: Highlight satin folding, a zipper closing, leather grain, bead movement, chain contact, stitching, or tissue wrapping. Production problem: Macro reshoots require controlled light and sound, while fast AI motion can distort seams or hardware. APOB workflow: Use a clean product frame, request one slow tactile action, lock the camera, and specify a restrained cloth, metal, or paper sound. Format: 9:16 detail Reel or a 16:9 visual interlude in a campaign edit. Limitation: Treat the clip as a stylized product visualization unless it faithfully represents the real material.
Beverage and food concept testing
Who it is for: Food creators and beverage marketing teams. Task: Test the visual hook of ice cracking, a can opening, tea pouring, sugar crust breaking, pastry layers separating, or a carbonated surface. Production problem: Liquids, utensils, hands, bites, and heat are difficult generative-physics cases, and audio can drift away from the contact point. APOB workflow: Prototype one action per clip, avoid a full recipe sequence, and regenerate when pouring direction, finger count, utensil contact, or sound timing fails. Format: Concept board, social test, or pre-production reference. Limitation: Do not present an impossible or synthetic food shot as evidence of actual taste, ingredients, preparation, or product performance.
Ambient scenes for a longer human-edited video
Who it is for: A YouTube creator assembling a study, reading, or atmospheric background video. Task: Produce short visual inserts such as rain beading on a window, tea steam under lamplight, a turning page, or embers settling. Production problem: AI generators are better suited to short clips than hours of stable, nonrepeating ambience. APOB workflow: Generate and approve several restrained clips, then combine them with licensed/original longer audio and human editing. Format: 16:9 YouTube video. YouTube reported in its 2026 CEO letter that Shorts average 200 billion daily views, but that scale is a distribution context—not a promise that any ASMR clip will perform. Limitation: Check loop cuts, audio licensing, repetition, and monetization originality before publishing a long compilation.
Short-form distribution is also material for product and fashion scenarios: Meta reported in September 2025 that Instagram had 3 billion monthly active users and Reels were reshared more than 4.5 billion times per day across Meta platforms. These figures explain why teams test vertical formats; they do not prove that AI ASMR improves reach or sales.
Production route | Best fit | Main tradeoff |
|---|---|---|
Fully generated AI ASMR | Impossible materials, fictional hosts, fast visual concept tests, or products that can be anchored to an approved image | Motion, anatomy, product details, and sound timing can vary; authenticity depends on clear disclosure and intentional editing |
AI-assisted ASMR | Generate or animate the visual in APOB, then add original or licensed Foley, ambience, whisper, or spatial mix in an editor | Adds post-production, but gives more control when the exact sound matters more than native generation speed |
Traditional human-recorded ASMR | Personal-attention formats, real product demonstrations, binaural microphone work, or scenes where physical truth is essential | Requires a performer, props, set, recording gear, releases, and reshoots, but offers the strongest control over genuine touch and sound |
Use fully generated video for openly synthetic scenes, AI-assisted production for precise sound design, and traditional recording when human presence or real-world product evidence is the reason viewers watch.
Audio and motion can be directed together Native Audio lets one prompt connect the contact point, material, action, and sound. This reduces the need to search for a generic effect after generation, although the result still needs timing review.
Impossible materials become testable concepts Glass fruit, jelly keyboards, cloud bread, or mineral candy can be prototyped without fabricating a physical prop. This is useful for entertainment concepts that are clearly synthetic.
A first frame can anchor a product or host Image to Video gives marketers and character creators more control than asking text alone to reconstruct the same subject. A high-quality source still matters.
One idea can produce channel-specific tests Creators can plan vertical trigger clips and wide ambient inserts without filming separate sets. The source ratio and current output controls should be chosen before generation.
The workflow connects to characters, Elements, and voices A recurring fictional host, product Element, or authorized/designed voice can support a series where the idea requires consistency rather than anonymous one-offs.
Physics and anatomy can fail Hands, blades, liquids, bites, folds, reflections, and product labels remain difficult. Regenerate or simplify the action instead of hiding a misleading frame behind fast editing.
Sound may be plausible but mistimed A click can arrive before contact, a cut can produce the wrong material sound, or the model can add speech or music. Always unmute and review; explicitly exclude unwanted audio.
Short clips still need editing for long-form ASMR A generated shot is not automatically a seamless loop or an hour-long ambience video. Longer projects need human selection, trimming, pacing, and licensed/original audio management.
Some ASMR audiences reject synthetic content For many viewers, human presence, effort, spontaneity, and trust are part of the trigger. AI is better suited to openly synthetic satisfying scenes, previsualization, or disclosed fictional formats than to impersonating a human ASMR creator.
Credits, watermarks, and privacy vary by plan Video cost depends on duration, quality, resolution, and options. Free-plan downloads carry a watermark, and private models/content are paid features. Check the current Generate cost and visibility before a client run.
No Credit Card Required
Answers about native audio, prompts, credits, watermarks, character consistency, commercial use, privacy, and disclosure.
What is AI ASMR, and how is it different from traditional ASMR?
AI ASMR uses generated or animated audio, video, or both to simulate triggers such as cutting, tapping, pouring, brushing, crinkling, whispering, or rain. Traditional ASMR records a real performer, object, space, and microphone. AI is useful for impossible or fictional scenes; human-recorded ASMR is usually better when authentic presence or precise binaural sound is the main appeal.
Can AI make ASMR videos with sound?
Yes. Current APOB Text to Video documentation says Native Audio is available at every Text to Video quality level; Ultra S requires it and does not let the user switch it off. Image to Video also exposes Native Audio in supported setups. Listen to the complete result because generated sound identity and contact timing can still vary.
How do I create an AI ASMR video in APOB?
Sign in, open a Personal or available Community model, and choose Video → Text to video. Keep Native Audio on, describe one subject, material, action, camera setup, and sound cue, then choose the live aspect ratio, duration, resolution, and quality. Confirm the credit total on Generate. The job and completed result appear in the Create feed, while notifications report success or failure.
Should I use text-to-video or image-to-video for ASMR?
Use Text to Video for surreal objects or scenes where exact identity is not critical. Use Image to Video for a branded product, recurring host, or composition that must start from an approved visual. In Image to Video, the source image establishes the initial frame; supported setups can add a last frame or multiple shots, but the APOB Image to Video guide says those two controls cannot be used together. Fix hands, labels, and object geometry before animation.
Can I use an AI ASMR generator for free without signing up?
APOB currently offers a Nano free plan and says no credit card is required to start, but an account is needed to retain credits, generate, and save work. The live billing guide currently lists an 80-credit daily Nano grant. Some quality and duration combinations cost more than that grant—the August 31 walkthrough showed 1,000 credits for a five-second 720P Ultra S Text to Video test with Native Audio. Nano content is public, so do not use confidential client or product assets for a free-plan test.
How much does a video cost, and will it have a watermark?
Video cost depends on plan, duration, resolution, quality, and enabled controls. APOB shows the current per-second rate and total beside Generate; the dated walkthrough example is documented in Step 3 above. Nano downloads carry a watermark, and the content and download guide says upgrading does not remove a watermark from content made before subscribing. If a generation fails, the APOB troubleshooting guide says the credits are returned automatically.
Is APOB an AI ASMR app or a website?
APOB runs in a web browser, and its current Image to Video guide documents desktop and mobile layouts. This page does not claim a separate native mobile app. Check the official APOB site for current availability rather than downloading an unrelated app with a similar name.
Can APOB create whispered AI ASMR voices?
Supported audio-enabled Image to Video workflows can add a Voice model for a spoken line. APOB can design a voice from a text description or train one from an authorized 10–90 second recording; custom voice models currently require Macro or above. A generated voice is not a substitute for binaural recording or detailed Foley, and a real person's voice should never be cloned without explicit permission.
Does Native Audio create binaural or studio-mastered ASMR?
Native Audio generates sound with the clip, but APOB does not claim guaranteed binaural placement, frame-perfect Foley, or a finished studio master. If spatial sound, exact timing, noise control, or long ambience is central to the experience, export the approved visual and finish with original or licensed audio in an editor.
How can I make an AI ASMR clip loop smoothly?
Request one continuous action, a locked camera, and a stable background. Inspect the first and last frames, then trim and align the approved clip in an editor. Automatic replay in a viewer does not mean the exported file is a seamless loop.
How do I reduce distorted hands, tools, liquids, or products?
Start from a clean first frame when identity matters, simplify the action, show fewer hands and objects, and avoid simultaneous cutting, pouring, speech, and camera motion. The APOB troubleshooting guide recommends trying Ultra when Fast produces quality problems. Regenerate when core geometry fails; do not crop a misleading product or unsafe interaction into an apparently valid result.
Can I keep the same AI ASMR character across videos?
A Portrait Model, a consistent approved first frame, and supported Elements can improve continuity. Image to Video supports Elements in Ultra and Ultra S, while Text to Video supports them in Ultra S. These references do not guarantee identical faces, hands, clothing, or props in every frame, so keep movement restrained and compare every result with the reference.
Can I publish AI ASMR on TikTok, Instagram, and YouTube?
Yes. Use 9:16 for TikTok, Reels, and Shorts and 16:9 for wide YouTube projects when appropriate. Open the completed item in the Create feed, unmute it, and inspect the downloaded file before publishing. Follow each platform's current disclosure, originality, copyright, advertising, and monetization rules; a successful export does not guarantee platform eligibility or audience acceptance.
Can I use APOB AI ASMR videos commercially?
The APOB terms permit legitimate commercial and creative use and assign generated-content rights to the user as between the user and APOB, to the extent permitted by law. You still need rights to every input and consent for real faces or voices. Review current terms, releases, local law, and platform rules before a paid campaign.
Who owns my inputs, and is my AI ASMR content private?
You must own or have permission to submit product images, music, recordings, faces, and voices. Do not assume uploads are confidential: private models/content are paid features, and visibility is not the same as data-processing confidentiality. Review the selected setting, Privacy Policy, and terms before uploading sensitive assets.
Do I need to label an AI ASMR video as AI-generated?
Label realistic synthetic video or audio whenever the destination platform requires it, and disclose earlier when viewers could mistake a scene, person, voice, or product behavior for real footage. TikTok, YouTube, and Meta use separate label systems. Disclosure does not authorize impersonation, infringement, deceptive claims, or fake endorsements.
