All Tools

Minimax H3 Prompt Generator

Paste your rough video idea. Get a production-ready Minimax H3 prompt with timestamps, camera moves, lighting, and audio — optimized for Hailuo AI.

1. Your Video Idea

0/2000

2. Parameters

6s · 2–3 shots
4s10s15s

Free · No sign-up needed · Powered by Claude

Before & After Examples

Before

a dog running on the beach

After

00:00.000–00:05.000 A golden retriever sprints along a wet shoreline at sunset, paws kicking up sand and water droplets that catch the orange backlight. Slow-motion tracking shot follows from a low angle. Waves crash softly in the background, seagulls call overhead. Cinematic, shot on 35mm film with shallow depth of field.

Added: specific breed, time of day, camera angle, motion detail, audio, visual style

Before

coffee being poured

After

00:00.000–00:03.000 Extreme close-up of dark espresso pouring from a brass Turkish pot into a white ceramic cup. Steam curls upward in slow motion, backlit by warm morning light through a kitchen window. 00:03.000–00:06.000 Camera slowly pulls back to reveal a cozy café table with a croissant and open book. Soft jazz piano plays. Shallow depth of field, warm color grading, photorealistic.

Split into 2 shots with timestamps. Added: specific coffee type, vessel, environment, audio, camera movement

Before

woman walking in city

After

00:00.000–00:04.000 A woman in her late 20s wearing a tailored navy coat walks confidently down a rain-soaked Tokyo side street at night. Neon signs reflect in puddles around her feet. Steadicam tracking shot from the front, slight low angle. 00:04.000–00:07.000 She pauses under a red paper lantern, looks up, and smiles. Slow crane up reveals the dense cityscape of glowing signs and steam rising from food stalls. Ambient city sounds — distant traffic, muffled J-pop from a shop. Cinematic anamorphic, blue and orange color palette.

Added: specific city, weather, clothing, expression, camera moves, multi-shot structure, audio atmosphere

The 7 Dimensions of a Great H3 Prompt

Minimax H3 responds to specific, layered prompts. Each dimension you add makes the output more controllable.

1

Subject

Who or what is in the frame. Be specific about appearance, clothing, age, expression.

young womanelderly mangolden retrievermatte black sports car
2

Action & Motion

What the subject does. Use active verbs. Describe speed, direction, and physics.

walks slowly toward cameraspins and catchespours liquid into glassdrifts around corner
3

Scene & Environment

Where it happens. Include time of day, weather, background details, depth.

rooftop at golden hourrain-soaked Tokyo alleyminimalist white studiodense forest with fog
4

Visual Style

The overall look. Pick one style and commit — mixing styles confuses the model.

photorealisticcinematic film grain35mm analogStudio Ghibli anime
5

Camera Movement

How the camera moves. Use specific filmmaking terms. One move per shot works best.

slow dolly inorbit shot 360°crane up revealhandheld tracking
6

Lighting & Color

Light direction, quality, and color palette. Sets the entire mood.

golden hour backlightharsh top-down spotlightsoft diffused overcastneon pink and cyan
7

Audio & Sound Design

Background sounds, music style, ambient noise. H3 generates audio natively.

soft piano melodycity traffic ambiencethunder and rainupbeat electronic beat

Before & After Prompt Examples

See how vague prompts become production-ready H3 prompts with timestamps, camera direction, and audio cues.

Before

a dog running on the beach

After
00:00.000–00:05.000 A golden retriever sprints along a wet shoreline at sunset, paws kicking up sand and water droplets that catch the orange backlight. Slow-motion tracking shot follows from a low angle. Waves crash softly in the background, seagulls call overhead. Cinematic, shot on 35mm film with shallow depth of field.

Added: specific breed, time of day, camera angle, motion detail, audio, visual style

Before

coffee being poured

After
00:00.000–00:03.000 Extreme close-up of dark espresso pouring from a brass Turkish pot into a white ceramic cup. Steam curls upward in slow motion, backlit by warm morning light through a kitchen window.
00:03.000–00:06.000 Camera slowly pulls back to reveal a cozy café table with a croissant and open book. Soft jazz piano plays. Shallow depth of field, warm color grading, photorealistic.

Split into 2 shots with timestamps. Added: specific coffee type, vessel, environment, audio, camera movement

Before

woman walking in city

After
00:00.000–00:04.000 A woman in her late 20s wearing a tailored navy coat walks confidently down a rain-soaked Tokyo side street at night. Neon signs reflect in puddles around her feet. Steadicam tracking shot from the front, slight low angle.
00:04.000–00:07.000 She pauses under a red paper lantern, looks up, and smiles. Slow crane up reveals the dense cityscape of glowing signs and steam rising from food stalls. Ambient city sounds — distant traffic, muffled J-pop from a shop. Cinematic anamorphic, blue and orange color palette.

Added: specific city, weather, clothing, expression, camera moves, multi-shot structure, audio atmosphere

Before

product showcase of a sneaker

After
00:00.000–00:03.000 A matte white sneaker with neon green accents sits on a glossy black turntable, rotating slowly. Camera orbits at eye level. Hard studio spotlight from above creates dramatic shadows. Clean white background.
00:03.000–00:06.000 Crash zoom into the sole tread pattern, then rack focus to the knit texture of the upper. Subtle bass-heavy electronic pulse. Hyperrealistic product photography style, 8K detail.

Added: product details, turntable motion, camera orbit + crash zoom, studio lighting, music, style reference

Before

make a fantasy scene with a dragon

After
00:00.000–00:05.000 A massive obsidian dragon with iridescent scales perches on a cliff edge above a cloud-filled valley at dusk. It spreads its wings slowly — wingspan fills the frame. Camera crane up from below reveals the full scale. Volumetric fog swirls around its talons.
00:05.000–00:09.000 The dragon inhales, chest glowing orange, then exhales a torrent of blue fire into the sky. Camera tracks the flame upward. Deep rumbling roar echoes through the valley, fire crackles. Epic orchestral swell. Cinematic fantasy, inspired by Game of Thrones cinematography.

Added: dragon appearance details, specific actions with physics, multi-shot, environmental depth, sound design, style reference

Camera & Style Cheat Sheet

🎥 Camera Moves

static shotslow dolly inslow dolly outdolly zoomtracking shothandheld trackingorbit shotorbit 360°crane upcrane downcrane revealpan leftpan rightwhip pantilt uptilt downzoom inzoom out

💡 Lighting

golden hourblue hourmagic hourhard sunlightsoft diffused lightovercast flat lightbacklight silhouetterim lighthair lightneon glowpractical lightscandlelightspotlighttop-down lightunder-lightingvolumetric light raysgod rayslens flare

🎨 Visual Styles

photorealistichyperrealisticcinematic35mm filmanamorphicIMAXStudio GhibliPixar 3Danimewatercoloroil paintingcharcoal sketchclaymationstop motionpaper cut-outcyberpunksteampunkvaporwave

Shot Budget & Timestamp Rules

4–6s
1–2 shots

Single continuous shot works best. If 2 shots, keep each 2-3s.

7–10s
2–3 shots

Sweet spot for storytelling. 3-4s per shot.

11–15s
3–5 shots

Plan carefully. More shots = more chance of inconsistency.

Timestamp Format

00:00.000–00:03.500 [Shot 1 description]
00:03.500–00:06.000 [Shot 2 description]
  • Format: MM:SS.mmm (minutes:seconds.milliseconds)
  • Each range = one continuous shot. No overlapping ranges.
  • Keep shots between 2–5 seconds each.
  • First shot starts at 00:00.000

Common Mistakes to Avoid

Don't use soft dissolve transitions — H3 handles cuts automatically

Don't request watermarks, logos, or text overlays — they render poorly

Don't describe "slideshow" pacing — each shot should have continuous motion

Don't mix incompatible styles (e.g. "photorealistic anime") — pick one

Don't write shots shorter than 1.5 seconds — too fast for the model

Don't specify exact frame counts or FPS — the model controls timing

Don't include meta-instructions like "make it viral" or "high quality" — describe what you see and hear instead

Don't overlap timestamp ranges — each range must be sequential

Frequently Asked Questions

What is Minimax H3 (Hailuo AI)?+
Minimax H3 is a video generation AI model developed by MiniMax, available through the Hailuo AI platform. It generates videos from text prompts (T2VA), images (I2VA), or reference images (Ref2VA) with native audio generation — meaning it creates both video and sound from your prompt.
How do I write timestamps in H3 prompts?+
Use the MM:SS.mmm format for timestamp ranges. Each range represents one shot — e.g. "00:00.000–00:03.500" for a 3.5-second shot. Ranges must be sequential (no overlapping), each shot should be 2-5 seconds, and the final range should end at your total video duration.
What aspect ratios does Minimax H3 support?+
Minimax H3 supports three aspect ratios: 16:9 (landscape, best for YouTube and desktop), 9:16 (vertical, best for TikTok, Reels, and Shorts), and 1:1 (square, best for Instagram feed and ads).
How do I add dialogue to H3 video prompts?+
Use the <d>[English] Your dialogue text here</d> syntax inside your shot description. Keep dialogue short — 1-2 sentences per shot maximum. H3 generates speech natively from this tag, so you don't need a separate TTS step.
What is the optimal video duration for H3?+
H3 supports 4-15 second videos. The sweet spot is 6-10 seconds. For 4-6s, use 1-2 shots. For 7-10s, use 2-3 shots. For 11-15s, use 3-5 shots. Longer videos with more shots have a higher chance of visual inconsistency between shots.
Is this prompt generator free?+
Yes, the prompt optimizer is completely free. You need to sign in with an inReels account (also free), but no credits are charged. You can optimize as many prompts as you want and use them on any platform that supports Minimax H3 / Hailuo AI.
What is the difference between T2VA, I2VA, and Ref2VA modes?+
T2VA (Text-to-Video+Audio) generates video purely from a text prompt. I2VA (Image-to-Video+Audio) animates a still image — your uploaded image becomes the starting frame. Ref2VA (Reference-to-Video+Audio) uses a reference image for style or subject guidance but generates a new scene, not a direct animation of the reference.

Write Prompts That Get Results

Stop guessing. Paste your idea, get a production-ready H3 prompt in seconds.

Try inReels Free — No Credit Card

Free prompt optimizer + 15 free credits for video generation

Read the full Minimax H3 prompt engineering guide →

Questions? Chat with us

We typically reply instantly