MiniMax H3 API for Developers
MiniMax H3 API
Generate AI videos with synchronized native audio using MiniMax H3 via API. Text-to-video and image-to-video, clips up to 30 seconds, no GPU setup required.
Free credits on signup
Transparent per-credit pricing
No watermarks on paid plans
Subscription or usage-based
Auto-scaling
Neon ramen
Dance alley
Desert mechanic
Glass blower
Puppet workshop
Storm porch
Barista pour
Trusted by teams at
Everything You Need to Ship MiniMax H3
Video generation with native audio and long-form clip support.
Native Audio
Audio is generated with the video in a single pass, synchronized to the action. No separate TTS or sound pipeline.
Image-to-Video
Animate a start-frame image with a text prompt, or generate straight from text on the text-to-video endpoint.
Up to 30 Seconds
Generate clips up to 30 seconds in one request — double the ceiling of most video models.
1080p Output
Render at 480p, 720p, or 1080p in 16:9, 9:16, or 1:1 to match the surface you're shipping to.
Auto-Scaling Infrastructure
Scales compute to match demand automatically. No GPU provisioning or cold starts.
Commercial-Ready
Paid plans support non-watermarked outputs with commercial licensing for production delivery.
Integrate in 3 Steps
Add AI video generation with sound to your app in minutes.
1
Get Your API Key
Sign up free and generate an API key from the Developer dashboard. You get free credits to start.
2
Send Your Request
Pass a text prompt or start-frame image to the image-to-video endpoint and select the minimax-h3 model. Set resolution and duration to taste.
3
Get Results & Scale
Receive the generated video via response or webhook. Deploy to production — infrastructure scales automatically.
Code Examples
MiniMax H3 in a few lines of code. Install an SDK or use the REST API directly.
from magic_hour import Client
from os import getenv
client = Client(token=getenv("API_TOKEN"))
res = client.v1.image_to_video.generate(
assets={"image_file_path": "/path/to/first-frame.png"},
end_seconds=10,
model="minimax-h3",
name="MiniMax H3 Video",
resolution="720p",
style={"prompt": "A ramen stall at night, steam rising, rain on the awning"},
)Estimate Your API Costs
Build with model-specific APIs for video, image generation, and image editing. Compare inputs, examples, and capabilities across Magic Hour's supported models.
Pricing tier options
Seconds of Video Per Month
100 seconds
seconds
Recommended plan
Pro Annual
$300
/year
300,000 credits per year
Up to 3.5 hours of video
Up to 30,000 images
PRO ANNUAL covers your usage
- 24,000credits / month
- 25,000credits / month
- 1,000credits / month
Estimated usage
Included
Remaining headroom
TOTAL BILL
$300
/year
SDKs for Every Stack
Official open-source SDKs with full TypeScript/type support. Install and start building.
MiniMax H3 API Pricing
Subscription plans for standard usage.
Prefer usage-based pricing?
Email us or message us in Intercom and we will set up your usage-based pricing.
Built for Builders
See how developers use the MiniMax H3 API to power real products.
Social Media & Short-Form
Ship TikToks, Reels, and Shorts that arrive with sound already attached — no audio pass required.
Ad Creative & Marketing
Generate campaign videos with ambient sound and consistent framing across a set of variants.
Product Demos
Create product walkthroughs long enough to show a full interaction rather than a fragment.
Scene Pre-Visualization
Test how a scene plays with sound before committing to a shoot or a full production pass.
Batch Variant Generation
Generate many takes of the same concept at 480p, then re-render the winner at 1080p.
Content Automation
Power video pipelines for editorial, UGC, and batch content workflows at scale.
Try before integrating
Try MiniMax H3 Free — No Code Required
Test quality first with the free tool, then plug the same output quality into your app with the API when you are ready.
MiniMax H3 API — Frequently Asked Questions
Magic Hour's MiniMax H3 API lets developers generate AI videos with native audio using MiniMax's Hailuo-family H3 model via REST API. It supports text-to-video and image-to-video workflows with SDKs for Python, Node.js, Go, and Rust.
Sign up at magichour.ai, generate an API key from the Developer dashboard, and install an SDK or call the REST API directly with model set to minimax-h3.
MiniMax H3 accepts a text prompt on its own, or a text prompt plus a start-frame image on the image-to-video endpoint. Supported image formats include JPG, PNG, and WebP.
No. MiniMax H3 generates from a prompt and an optional start frame. If you need to pin the last frame of a clip, use a model that supports end frames such as LTX 2.3 or Kling 3.0.
MiniMax H3 generates videos up to 30 seconds per request at 24fps, longer than most models in the catalog.
MiniMax H3 outputs at 480p, 720p, and 1080p in 16:9, 9:16, and 1:1 aspect ratios.
Yes. MiniMax H3 generates audio natively alongside the video, synchronized to the action. No separate audio pipeline or post-production layering needed.
MiniMax H3 costs 24 credits per second at 480p, 48 at 720p, and 72 at 1080p, so a 10-second 720p clip runs 480 credits. Magic Hour offers subscription plans starting at $19/month (Creator), $39/month (Pro), and $99/month (Business), each including a monthly credit allotment. You can also contact us to set up usage-based (pay-as-you-go) pricing.
Sign up at magichour.ai/developer — no credit card required — to get an API key instantly, along with free credits. The key works across all Magic Hour API endpoints and our Python, Node.js, Go, and Rust SDKs.
Signing up is free and includes credits you can spend on MiniMax H3 at 480p. Rendering at 720p or 1080p requires a paid plan (Creator, Pro, or Business), which also unlocks higher credit limits, concurrency, and commercial use rights.