Quick Answer
These 15 viral AI prompts cover the formats actually spreading in 2026: AI microdramas, talking animals, ASMR clips, cinematic images, hooks, faceless explainers and product ads. Each one is copy-paste ready for Veo, Kling, Runway, Midjourney, Nano Banana, ChatGPT or Claude. A prompt removes the technical barrier, but it never guarantees virality. The idea, the timing and the audience still decide. For the originals that first proved the format, see these copy-paste AI prompts that went viral.
What Is Actually Going Viral in 2026?
Trends change within weeks, so learn the structure, not one viral video. Six repeatable TikTok AI formats kept appearing through 2026: AI microdramas (short vertical episodes with a cliffhanger, usually 2 to 4 minutes), surreal character memes, talking animals and objects, ASMR-style satisfying clips, cinematic B-roll under a voiceover, and "what if" alternate-history clips.
AI baby reactions were one of the fastest-growing formats of early 2026. The joke is an infant face with burned-out adult behaviour. "Baby with Mustache AI" spread from 21 October 2026, and "Bonito Bonito AI Baby" from 23 October 2026. Recurring characters build followings; one-tap effects do not.
The scale possible is real, and so is the platform risk. Fruit Love Island, an AI series about anthropomorphic fruit in a dating show, launched 13 March 2026 and reached 3 million followers and 300 million views in nine days. The creator stopped production on 28 March 2026 after TikTok began removing episodes. That is the honest picture: huge reach is possible, and the platform can pull the plug.
How to Use These Prompts
Copy the prompt, paste it into the tool named, and change only the bracketed parts like [TOPIC] or [AUDIENCE]. A few rules make the difference between a clip that works and credits wasted.
For Veo, follow Google's formula: cinematography + subject + action + context + style + audio. Put dialogue in quotes, mark sound effects with "SFX:" and background sound with "Ambient noise:". Keep dialogue to 8 to 12 words per 8-second clip, and keep the whole prompt between 40 and 120 words. Set 9:16 in the parameters, not in the prose. One shot, one action per clip.
For Kling, keep prompts short with one clear action. For Runway image-to-video, always start from an uploaded photo. For Midjourney, use only current parameters and leave the version flag off so it uses the current default model. Budget for two or three takes per finished clip. Veo is not deterministic: the same prompt gives different clips, and on credit-based tools every failed take costs the same as a good one. Finish every video in an editor like CapCut, free at 1080p with no watermark as long as no Pro item is in the project.
Viral Short-Video Prompts
Prompt 1 — AI Microdrama Scene (Veo 3.1)
Slow push-in, medium close-up, shallow depth of field. A young woman in a rain-soaked coat stands under a flickering streetlamp in a narrow alley at night, holding a torn photograph, tears mixing with rain on her face. She looks straight into the camera and says, "You told me you would come back." Neon signs glow pink and blue behind her, cinematic teal-and-orange grade, film grain. SFX: distant thunder, rain pattering on pavement. Ambient noise: low city hum.
Tool: Veo 3.1 in Flow. Why it works: one emotional scene plus one short line equals one 8-second clip, and a series of these becomes a microdrama episode with a cliffhanger. Pro tip: the dialogue is 8 words, inside the 8 to 12 that fit an 8-second clip. End each clip on the line, not after it.
Prompt 2 — Talking Animal at Work (Kling)
A golden retriever in a tiny business suit sits upright behind a wooden office desk, typing on a laptop with its paws, then looks up at the camera with an exasperated sigh and shakes its head. Bright modern office, sunlight through large windows, shallow depth of field, photorealistic, smooth natural motion, vertical video.
Tool: Kling. Why it works: talking animals are a top format of 2026, and the joke is contrast: real animal behaviour in a human job. "Dog Bartender AI" and "Kung Fu Cat AI" were spreading in early November 2026. Pro tip: works best as image-to-video from a start image of the dog in the suit. A 5-second 1080p clip with audio costs about 60 credits on Kling 3.0.
Prompt 3 — ASMR Satisfying Clip (Veo 3.1)
Extreme macro close-up, static camera, shallow depth of field. A sharp knife slowly slices through a glossy, jewel-like glass strawberry on a dark slate surface, the cut revealing sparkling red crystal layers inside. Soft studio lighting, slow deliberate motion, hyper-detailed textures. SFX: crisp glass slicing sound, tiny crystal chime. Ambient noise: near silence.
Tool: Veo 3.1 in Flow. Why it works: ASMR-style satisfying clips are built for sound and loops, and they need no dialogue, so there is no lip-sync risk. Pro tip: one shot, one action, no dialogue is the safest Veo prompt type. A static camera avoids motion artefacts.
Viral Image Prompts
Prompt 4 — Cinematic Portrait (Midjourney)
A cinematic portrait of a weathered fisherman in his sixties, salt-stained skin, grey stubble, wool sweater, standing on a foggy harbour pier at dawn, holding a coiled rope, dramatic side lighting, shallow depth of field, muted teal and amber colour grade, film grain, 35mm lens look, photorealistic --ar 4:5 --stylize 250
Tool: Midjourney. Why it works: cinematic portraits are the most-shared AI image format on Instagram, and the movie-poster look is what stops the scroll. Pro tip: only current parameters are used here. --ar sets the aspect ratio, --stylize 250 adds a painterly mood. Leave the version flag off so it uses the current default model. --test and --upbeta are obsolete since version 3, never use them. Midjourney has no free tier: Basic starts at $10 a month. For more still-image ideas, see these AI photo prompts.
Prompt 5 — History Reenactment Photo Edit (Nano Banana, Free)
Edit this photo into a 1920s black-and-white street photograph: keep the same person, same pose and same framing, but change the setting to a cobblestone street in 1920s Paris, add a vintage wool coat and flat cap, period cars and shop signs in the background, film grain, slight blur at the edges, sepia tone, authentic early-photography look.
Tool: Nano Banana in the Gemini app (free). Why it works: the contrast between a modern person and an old world is instantly readable, and "what if" formats are a top trend of 2026. Pro tip: this is an edit prompt for an uploaded photo, not a text-to-image prompt. Upload your own photo first, then paste it. Nano Banana keeps the person consistent while changing the scene, which is why it is the best free tool for this.
Prompt 6 — Luxury Product Hero Shot (Midjourney)
A luxury product photograph of a matte black wireless earbuds case floating above a reflective dark surface, dramatic rim lighting, soft mist, water droplets on the case, deep shadows, premium advertising style, ultra-detailed, studio lighting --ar 4:5 --stylize 100
Tool: Midjourney. Why it works: floating-product hero shots are the most reused AI ad image format, working as thumbnails, ad creatives and store listings. Pro tip: --stylize 100, the default, keeps the prompt literal so the product stays accurate. High stylize values distort product details. Swap the earbuds case for your own product. If you need the exact real product, use Nano Banana or ChatGPT Images with a photo of the real product instead, because Midjourney invents details.
Viral Hooks & Captions
Prompt 7 — Hook Generator (ChatGPT or Claude)
Act as a viral short-form scriptwriter with 10 years of experience on TikTok and YouTube Shorts. Write 10 opening hooks (under 12 words each) for a video about [TOPIC]. Rules: every hook must create curiosity or tension in the first 3 seconds, no clickbait that the video cannot deliver, mix 3 question hooks, 3 bold-claim hooks, 2 story hooks and 2 "mistake" hooks. For each hook, add one line explaining why it stops the scroll. My audience: [AUDIENCE]. My tone: [FUNNY / BOLD / CALM].
Tool: ChatGPT or Claude, free tiers work. Why it works: the hook decides whether a short gets watched or swiped, and this prompt is reusable for every video you make. Pro tip: the "no clickbait the video cannot deliver" rule matters. A hook that lies kills retention and trust.
Prompt 8 — Comment-Bait Caption (ChatGPT or Claude)
Turn the idea below into a short Instagram/TikTok caption that gets comments. Start with a one-line hook, then 3 short lines that tell a relatable mini-story, then end with ONE specific question that is easy to answer in a comment (not "what do you think?"). Keep it under 80 words. Tone: [CASUAL / BOLD / FUNNY]. No hashtags in the body, no emojis in the first line. My idea: [YOUR IDEA]. My niche: [YOUR NICHE].
Tool: ChatGPT or Claude. Why it works: comments are the strongest early engagement signal for short-form ranking, and a specific, easy question gets far more replies than a generic one. Pro tip: edit the output in your own voice before posting. AI captions read as AI captions until a human touches them.
Prompt 9 — Viral Text Post Writer (ChatGPT or Claude)
Turn my rough idea below into a viral text post for [X / LINKEDIN]. Start with a scroll-stopping first line under 12 words. Use short lines, one idea per line, no paragraph longer than 2 lines. Put a surprising fact or number in the middle. End with a question that invites replies. Keep the tone [BOLD / HELPFUL / FUNNY]. No hashtags. My rough idea: [YOUR IDEA].
Tool: ChatGPT or Claude. Why it works: the short-line, one-idea-per-line structure is the format that performs best on X and LinkedIn. Pro tip: check any "surprising fact or number" the AI invents before posting. AI chatbots fabricate statistics.
Faceless Channel Prompts
Prompt 10 — Faceless Explainer Script + B-Roll Shot List (ChatGPT or Claude)
Act as a YouTube scriptwriter for a faceless channel. Write a 60-second script about [TOPIC] for a viewer who knows nothing about it. Structure: hook in the first 5 seconds (a surprising fact or question), then 3 clear points, each with one example, then a one-line takeaway and a subscribe CTA. Write it in simple spoken English, short sentences, no jargon. After the script, list 6 B-roll shot descriptions (one per 10 seconds) I can generate with an AI video tool: each shot is one sentence describing subject, action and camera move. My audience: [AUDIENCE].
Tool: ChatGPT or Claude for the script, then Veo or Kling for the shots. Why it works: faceless explainers are the lowest-barrier YouTube format: no camera, no face, just a script, a voiceover and AI B-roll. The shot list is what makes this better than a plain "write me a script" prompt, because it bridges straight to video generation. If you are building a whole channel around this format, these 100 AI prompts for YouTube cover titles, scripts and growth. Pro tip: a 60-second video needs roughly 6 to 8 clips. Paste the B-roll descriptions one by one into Veo (8-second clips) or Kling (5 to 10 seconds).
Prompt 11 — "What If" Alternate History (Veo 3.1)
Slow aerial drone shot descending over a bustling Roman forum at golden hour, marble columns, crowds in togas, market stalls with bread and pottery, a steam-powered chariot rolling past in the foreground. Warm sunlight, long shadows, cinematic epic scale, photorealistic, smooth camera movement. Ambient noise: crowd murmur, distant hammering, cart wheels on stone.
Tool: Veo 3.1 in Flow. Why it works: one impossible detail (steam power in ancient Rome) inside an otherwise realistic scene is the whole hook. Pro tip: drone moves are the riskiest Veo camera move. One October 2026 test found a drone-rise prompt partly failed. Keep the move slow and simple. If it fails, regenerate, or drop the drone and use a static wide shot.
Prompt 12 — AI Baby With Adult Attitude (Veo 3.1)
Static medium shot, eye level, shallow depth of field. A baby in a tiny business suit sits in a high chair at a kitchen table, holding a miniature coffee mug, and looks at the camera with tired adult eyes and says, "Another Monday, another meeting." Warm morning light through a window, photorealistic, deadpan comic timing. SFX: soft coffee sip, quiet sigh.
Tool: Veo 3.1 in Flow. Why it works: the joke is the contrast between an infant face and burned-out adult behaviour, one of the fastest-growing formats of early 2026. Pro tip: the dialogue is 5 words, so it fits an 8-second clip easily. Some generators restrict realistic images of children and may refuse baby-face prompts. If Veo refuses, use a stylised version, like a clay figurine or cartoon baby, instead. Never use a real child's photo as the input.
Viral Product-Ad Prompts
Prompt 13 — Product Hero Ad (Veo 3.1)
Slow dolly-in, extreme close-up, shallow depth of field. A sleek matte black smartwatch rotates slowly on a dark reflective pedestal, its screen glowing with a bright fitness display, a single water droplet sliding down the glass. Dramatic rim lighting, deep shadows, premium commercial look, hyper-detailed. SFX: soft electronic chime, low whoosh as the watch turns.
Tool: Veo 3.1 in Flow. Why it works: the rotating hero shot is the standard AI product-ad opener. It shows the product, the screen and one premium detail, the water droplet, in a single 8-second clip. Pro tip: one shot, one action, no dialogue, so no lip-sync risk. Swap the product and the screen content for your own. For a real product, generate from a photo of the actual product instead of text alone.
Prompt 14 — Before/After Transformation (Runway Gen-4)
Use the uploaded photo as the first frame. The camera slowly pulls back as the dull, cluttered living room transforms into a bright modern space: the old sofa is replaced by a clean white couch, warm lamps switch on, plants appear on the shelves, sunlight streams through the window. Smooth continuous motion, photorealistic, interior-design commercial style, 5 seconds.
Tool: Runway Gen-4, image-to-video. Why it works: before/after transformations are the highest-converting ad format for home, beauty and decor products. The change is the ad. Pro tip: this prompt must start from an uploaded photo, the "before" room. A 5-second clip costs 60 credits on full Gen-4, or 25 on Gen-4 Turbo. Do not write this as a text-only prompt. Gen-4 needs the start image.
Prompt 15 — UGC-Style Testimonial Ad (Veo 3.1)
Handheld selfie-style medium shot, slight natural camera shake, natural window light. A cheerful woman in her thirties holds up a small amber glass skincare bottle, smiles at the camera and says, "Three weeks, and my skin has never looked better." Bright modern bathroom background, authentic phone-video look, vertical framing. Ambient noise: faint room tone.
Tool: Veo 3.1 in Flow. Why it works: UGC-style ads outperform polished commercials on TikTok and Reels because they look like a real person's video, not an ad. Pro tip: the dialogue is 9 words, fitting 8 seconds. Keep the shake "slight" or the model overdoes it. Honesty rule: a testimonial line is a template. Use your product's real result, or clearly frame it as a demo. Invented testimonial claims are deceptive advertising. Several platforms also require AI-generated ads to be labelled as AI.
Which Tool Should You Use?
| Tool | Status Nov 2026 | Free tier | Paid price | Best for in this post |
|---|---|---|---|---|
| Veo 3.1 (Flow) | Live, current model | 50 credits/day, no card | Plus $4.99/mo; Pro $19.99; Ultra $99.99/$199.99. Per 8s clip: Lite 10 credits, Fast 20, Quality 100 | Prompts 1, 3, 11, 12, 13, 15 — dialogue + sound in one generation |
| Runway Gen-4.5 | Live, current flagship | 125 credits one-time only | Standard $15/mo ($12 annual); Pro $35 ($28); Max $95 ($76). 5s clip: 60 credits, 25 on Turbo | Prompt 14 — image-to-video transformations |
| Kling 3.0 | Live | Monthly free credits, amount varies | Standard $8.80/mo ($6.99 first month); Pro $32.56; Premier $80.96 | Prompt 2 — talking animals, natural motion |
| Pika | Live | $0 with 0 monthly credits | Starter $10/mo ($8 annual); Creator $35 ($28); Fancy from $95 ($76) | Short clips where you want a model menu |
| Midjourney | Live | None — no free tier or trial | Basic $10/mo ($8 annual); Standard $30 ($24); Pro $60 ($48); Mega $120 ($96) | Prompts 4, 6 — cinematic images, product shots |
| Nano Banana | Live, free in the Gemini app | Free in Gemini app (Nano Banana 2) | Plus $4.99/mo; Pro $19.99; Ultra $99.99/$199.99 | Prompt 5 — photo edits that keep the person |
| ChatGPT Images | Live | Limited image creation, a few a day (third-party reports) | Go from $8/mo; Plus $20; Pro from $100 | Prompts 7–10 (text) + image edits |
| Claude | Live | Free, no card | Pro $20/mo ($17 annual); Max $100; Ultra $200 | Prompts 7–10 (text) — long structured scripts |
| CapCut | Live, the editor | Free, 1080p export, no watermark if no Pro items used | Pro $19.99/mo; Ultra $79/mo | Editing + captions for every video prompt |
One tool is deliberately missing from this list. Sora is dead. OpenAI shut the Sora app on 26 April 2026 and removed the Videos API and Sora 2 model aliases on 24 September 2026. Any guide, prompt pack or tool list that still names Sora as a live option is stale. The live video tools are Veo, Runway, Kling and Pika.
The Honest Truth About Going Viral
A prompt does not guarantee virality. No prompt, tool or format can promise views. What a good prompt does is remove the technical barrier. The idea, the timing and the audience still decide.
Results vary by tool version and by run. Veo is not deterministic, so the same prompt gives different clips. Budget for two or three takes per finished clip as practical advice, and remember that on credit-based tools every failed take costs the same as a good one.
Some tools need a paid plan for the best output. Midjourney has no free tier at all. Veo Quality clips need a paid plan, because free daily credits never buy one. Runway's free credits are one-time, not monthly.
AI video still fails. Hands, faces in busy scenes and complex camera moves break most often. One October 2026 test found a drone-rise prompt partly failed and a three-shot timestamp prompt failed outright. Keep prompts to one shot and one or two actions.
AI chatbots invent facts, so any statistic or "surprising fact" in a script must be verified before posting. AI-generated ads must be labelled where platforms require it, and testimonial-style ads must use real results.
Key Takeaways
- Veo 3.1 is the best video tool in this list because it generates dialogue, sound effects and ambience in the same clip. Six of the 15 prompts target it.
- The best free routes are Kling's monthly credits, Veo's 50 free daily credits, and Nano Banana free in the Gemini app. Midjourney has no free tier at all.
- One shot, one action per clip. Budget two or three takes, and finish every video in an editor like CapCut.
- Sora is discontinued and must not be used. Any guide still listing it is stale.
- No prompt guarantees views. Trends expire in weeks, so learn the structures, not one viral video.
Frequently Asked Questions
Do these prompts work on free tools?
Most of them, yes. Veo gives 50 free credits a day, Kling gives a monthly free allowance, Nano Banana is free in the Gemini app, and ChatGPT and Claude have free tiers for the text prompts. Midjourney needs $10 a month minimum, Veo Quality clips need a paid plan, and Runway's free credits are one-time only.
Which AI video tool is best for beginners?
Veo 3.1, because it generates dialogue, sound effects and ambience in one generation, and the free daily credits cover Lite and Fast clips. Kling is the other good start: the cheapest paid plan at $6.99 for the first month, and the best natural motion for faces and animals.
Why is Sora not on this list?
Sora is discontinued. OpenAI shut the Sora app on 26 April 2026 and removed the Videos API and Sora 2 model aliases on 24 September 2026. The live video tools are Veo, Runway, Kling and Pika.
How long should a viral AI video be?
One Veo clip is 8 seconds. Stitch 6 to 8 clips for a 60-second video. Microdrama episodes run 2 to 4 minutes. For Shorts and Reels, 15 to 60 seconds is the usual range. The format decides the length, not the tool.
Can I use these prompts for client work?
The prompts themselves are yours to use. The output is a different question: check each tool's commercial terms first. Kling's free tier does not allow commercial use, and Pika's commercial licence terms conflict between sources, so verify on the live page before selling client work made with them.
Do AI prompts guarantee views?
No. No prompt, tool or format can promise views. A good prompt removes the technical barrier. The idea, the timing, the audience and the first three seconds decide whether a video spreads.
Want to go deeper with video? Read our guide to creating viral AI videos with Veo, Runway & Pika, our 7 MidJourney prompts, and our 50 viral product ad prompts. For the tools behind the prompts, see our Vibe Coding tutorial and our AI meeting assistants guide. If you are building skills first, start with our list of 10 free AI certification courses, and grab the freebies in our Freebies Library.
Want more prompts? Grab the free 20 AI Video Prompts sampler, the full 50 AI Video Prompts pack ($9), or the free 50 Viral Product Ad Prompts pack from the Creator Vault store.

