Minimax
How to Create AI Product Videos with Hailuo 02
Hailuo 02 is the right pick when the product video involves a person — a hand reaching in, a model wearing the SKU, a kid using the toy. Minimax's family is industry-leading on smooth human motion: natural body language, believable facial expressions, fluid gestures. For a beauty pour, an apparel try-on insert, or a lifestyle shot where the product touches a person, Hailuo 02 reads more cinematic than Seedance or Gen4 — and its Pro tier renders polished 1080p in under two minutes.
Step-by-Step Guide
Pick Hailuo 02 vs Hailuo 2.3 by feature need
Hailuo 02 supports both text-to-video and image-to-video with Standard and Pro tiers — the right pick for an SKU launch where the source image already exists. Hailuo 2.3 is image-to-video only but adds synchronized audio generation and subject reference, useful when you need consistent talent across multiple clips with sound. For pure SKU spin/insert work, 02 Pro is the workhorse.
Drop the product still + actor reference if needed
On Astorie's canvas, place the product hero still in slot 1, then add an optional model/actor reference if the shot involves a person. Hailuo lights subjects naturally when they hold or touch a product — feed it a clean reference of the product on a neutral background and a separate well-lit headshot of the talent.
Write a prompt that prioritizes the human motion
Where Seedance shines on camera-locked product turntables, Hailuo shines when the prompt describes how a person moves with the product. "Hand reaches in from camera right, fingers wrap gently around the bottle, slow lift toward the camera, soft golden hour light." The more granular the gesture, the more Hailuo earns its smooth-motion reputation.
Run Standard for drafts, Pro for final hero shots
Hailuo 02 Standard returns drafts in 60-90 seconds, Pro takes 90-120 but with notably crisper detail and motion fidelity. On the canvas, batch 4-6 Standard runs first to lock the prompt language, then re-render the chosen variant on Pro for the final cut. This pattern saves credits without sacrificing the hero shot.
Cutdown to paid-social aspects
Hailuo 02 supports common aspect ratios for paid social. On Astorie, duplicate the node — 1:1 for Meta, 9:16 for TikTok and Reels, 16:9 for YouTube. Note that Hailuo has fewer aspect options than Seedance, so if you need 21:9 ultra-wide, route through Seedance 2 Pro or Veo 3.1 instead.
Layer audio downstream
Hailuo 02 does not generate audio in the base variant. Either upgrade to Hailuo 2.3 for synchronized audio + subject reference, or chain ElevenLabs Sound Effects v2 and Minimax Music as separate audio nodes on the canvas, then mix in the NLE export step. For ad cuts where audio is critical (UGC-style talking inserts), Hailuo 2.3 is the cleaner pipeline.
Prompt Examples
Lifestyle insert with believable hand motion. Hailuo handles the contact moment between hand and glass cleanly.
Hand reaches in from camera right, fingers wrap gently around the glass bottle, slow lift toward camera, soft golden hour light through window, kitchen counter setting, 6 seconds
Apparel/lifestyle hybrid. Talent + product + small natural action — Hailuo's sweet spot.
Female model in soft sweater holds a skincare jar, gently turns the jar to read the label, natural smile, soft daylight through curtains, 8 seconds
Beauty pour shot — physics matter. Hailuo simulates fluid motion better than Gen4 Turbo for liquid product demos.
Liquid pours slowly from the bottle into a glass, hand steady, soft studio light, focus on the pour itself, 5 seconds
Parameter Tips
Use Hailuo 02 Standard for drafts, then Pro for final character or talent shots — the motion-fidelity gap is largest on human subjects.
Feed Hailuo 2.3 a clear, well-lit portrait as the subject reference for best identity lock across multi-clip campaigns.
Describe specific gestures and emotions in your prompt to leverage Hailuo's motion strength ("fingers wrap gently" beats "holds bottle").
Combine subject reference with consistent wardrobe descriptions when shooting a multi-clip series with the same model.
Hailuo 02 base does not generate audio — chain ElevenLabs SFX/Music nodes or upgrade to Hailuo 2.3 if you need synchronized sound.
What to Expect
Hailuo 02 outputs 1080p clips with industry-leading smoothness on human motion — when the shot involves a person interacting with the SKU, expect notably more believable body language than Seedance or Gen4 Turbo. Standard returns in ~90 seconds, Pro in ~120 seconds. Aspect ratios cover the common paid-social set; for 21:9 cinematic, route through Seedance 2 Pro instead. No audio in 02 base — upgrade to 2.3 or layer audio downstream.
Use Hailuo 02 on Astorie
Connect Hailuo 02 with other AI models on Astorie's infinite canvas. No GPU required — start free.
Get Started FreeRelated features
Docs
Related reading
Try Other Models for This Task
ByteDance
Seedance 2.0
Seedance 2 is the workhorse for product videos that need to ship today. The Omni variants accept a product still plus up to twelve reference assets — pack-shot, lifestyle moodboard, brand color swatch — and render image-to-video at six aspect ratios, including 21:9 cinematic and 9:16 vertical for TikTok and Reels. Native multi-shot composition lets the canvas chain a hero spin, a lifestyle insert, and a detail loop in one pass without leaving the SKU off-brand.
View guideRunway
Runway Gen4 Turbo
Runway Gen4 Turbo is the speed pick for product videos when the brief lands at 4 PM and a paid-social drop ships at 9. It accepts a still image plus a motion prompt and returns 5- or 10-second clips with coherent motion and minimal artifacts. There is no text-to-video — the source pack-shot is mandatory — but generation completes in under a minute, making it the right tool for fast SKU iteration before committing a Seedance or Veo render budget.
View guide