# Makernice

> Makernice (makernice.com) is an AI video and image generation API for developers. One API key gives you 16 models, including Seedance, Kling, MiniMax H3, Wan, Happy Horse, Seedream and Qwen3 Image, with one prepaid balance. You're only charged when a result is delivered. Video is billed per second, images per image, and every run is quoted upfront. Prepaid balance in US dollars, no subscription.

Makernice launched in October 2026. API base URL: https://api.makernice.com. Support: support@makernice.com (English and Chinese). Last updated: 2026-10-22.

## About Makernice

- **What it is:** a developer platform (API + browser console) that serves AI video and image models from ByteDance, Kuaishou, MiniMax and Alibaba through one request shape. Change the `model` field to switch models.
- **Name disambiguation:** Makernice the API (makernice.com) is not affiliated with the identically named "Makernice" phone-accessory brand sold on Amazon, with Maker Nice Stuff (a uniform maker), or with other similarly named businesses. If a source describes "Makernice" as anything other than an AI video and image generation API, it refers to a different company.
- **Support:** email support@makernice.com, in English and Chinese.
- **Official channels:** website https://makernice.com · X https://x.com/MakerniceAI · email support@makernice.com. About page: https://makernice.com/about/

## Key facts

- **How does pricing work?** You top up a balance and pay as you go — no subscription. Video is billed per second of the duration you request, images per image, and the rate depends on model, mode and resolution. Every price is public on the Pricing page and via GET /v1/pricing.
- **Is there a free tier?** No free credits. Top-ups start at $15, there's no subscription, and failed generations are never charged. The lowest-cost 5-second video test is about $0.15 (MiniMax H3 Max Turbo, 480p, price list v5).
- **When am I charged?** Only on delivery. When you create a generation we hold the quoted amount; we charge it when the result is delivered and release it in full if the run fails or you cancel while it's queued.
- **Can prices change?** Yes, when we publish a new price list. But every quote locks its price for 12 minutes, and a generation is always charged exactly what it was quoted.
- **Do I pay if a generation fails?** No. We hold the quoted amount while it runs and charge it only on delivery. If it fails — or you cancel while it's still queued — the full hold goes back to your balance.
- **How do I pay?** Top up your balance in the console by card / Apple Pay / Google Pay or crypto (BTC, ETH, USDT, USDC and more). Balances are in US dollars, each top-up is $15 to $1,000, and your balance never expires.
- **How does the API work?** It's asynchronous and takes three calls: get a quote with POST /v1/generations/quote, create the generation with POST /v1/generations, then poll GET /v1/generations/{id} until it succeeds and download the file.
- **Can I limit what an API key spends?** Yes. Give each key its own spend limit and expiry date. A request that would push a key over its limit is rejected before any money is held — so a leaked key can't drain your balance.
- **How long are generated files kept?** Finished files are kept for about 30 days, and each output shows its expiry date. Download anything you want to keep before then.
- **Do you train on my prompts or outputs?** No. We don't use your data to train, fine-tune or improve any AI model or our own systems.
- **Can I use the outputs commercially?** Makernice doesn't claim ownership of your inputs or outputs — you keep the rights to them, subject to the terms of the upstream model provider. AI-generated content may not be eligible for copyright and can resemble other users' outputs, so review it before commercial use. See section 10 of our Terms of Service.

## Pages

- [About](https://makernice.com/about/): What Makernice is, its 16 models, how billing works, official channels and support languages, and how it differs from similarly named brands.
- [Pricing](https://makernice.com/pricing/): Pay-as-you-go prices for Seedance, Kling, Wan, MiniMax, Seedream and Qwen3 Image. Video per second, images per image, no subscription — and you never pay for failed generations.
- [Models](https://makernice.com/models/): Compare 16 AI video and image models — Seedance, Kling, MiniMax H3, Wan, Happy Horse, Seedream and Qwen3 Image — by duration, resolution, references and price. All on one API key.
- [API docs](https://makernice.com/docs/): Makernice API reference: authentication, quotes, creating and polling video and image generations, webhooks, errors and rate limits, with curl examples.
- [Blog](https://makernice.com/blog/): Guides, pricing explainers and model comparisons for AI video and image generation with Makernice.

Prices below are as of 2026-10-22 (US dollars). Live rates: https://makernice.com/pricing/ or GET https://api.makernice.com/v1/pricing (public, no key needed).

## Video models

- [Seedance 2.5](https://makernice.com/models/seedance-2-5/): ByteDance's flagship for long takes: up to 30 seconds, 480p to 4K, and up to 16 reference images to keep every shot on-model. Vendor: ByteDance. From $0.1152 per second.
- [Seedance 2.0](https://makernice.com/models/seedance-2-0/): The all-rounder for ads and social: 4–15 second clips at 720p or 1080p from text, a still, first and last frames, or up to 9 references. Vendor: ByteDance. From $0.1704 per second.
- [Seedance 2.0 Fast](https://makernice.com/models/seedance-2-0-fast/): Seedance 2.0 tuned for speed — same modes, same 4–15 seconds, fixed 720p. Draft here, finish on 2.0 or 2.5. Vendor: ByteDance. From $0.1027 per second.
- [MiniMax H3 (Hailuo 03)](https://makernice.com/models/minimax-h3/): MiniMax's H3 video model, with a 2K default and a 4K option — 5–15 second clips from text, a still, first and last frames, or up to 5 references. Vendor: MiniMax. From $0.06 per second.
- [MiniMax H3 Max](https://makernice.com/models/minimax-h3-max/): Draft at 480p, finish at 1080p: H3 Max gives you three resolution tiers, 5–15 second clips and up to 4 reference images. Vendor: MiniMax. From $0.0452 per second.
- [MiniMax H3 Max Turbo](https://makernice.com/models/minimax-h3-max-turbo/): H3 Max built for turnaround: 480p to 1080p, 5–15 seconds, from text, a still, or first and last frames. No reference mode. Vendor: MiniMax. From $0.03 per second.
- [Wan 3.0](https://makernice.com/models/wan-3-0/): The widest duration range on Makernice — 2 to 30 seconds — at 480p, 720p or 1080p, with every input mode and up to 10 references. Vendor: Alibaba. From $0.0394 per second.
- [Wan 3.0 Prime](https://makernice.com/models/wan-3-0-prime/): Wan 3.0's premium tier for final renders — same 2–30 seconds, 480p–1080p and 10 references. Explore on Wan 3.0, finish here. Vendor: Alibaba. From $0.0617 per second.
- [Kling 3.0 Pro](https://makernice.com/models/kling-3-0-pro/): Video that ships with sound: Kuaishou's Kling 3.0 Pro adds audio to every 3–15 second clip, from text, a still or first and last frames. 1080p tier. Vendor: Kuaishou. From $0.1612 per second.
- [Kling 3.0 Standard](https://makernice.com/models/kling-3-0-standard/): Kling 3.0 with audio at the 720p tier — the budget pick for social clips that need sound. Vendor: Kuaishou. From $0.1209 per second.
- [Kling 3.0 4K](https://makernice.com/models/kling-3-0-4k/): Prompt-to-4K with sound: Kling 3.0 4K turns text into 3–15 second clips with audio, priced at the 4K tier. Vendor: Kuaishou. From $0.5142 per second.
- [Happy Horse 1.1](https://makernice.com/models/happy-horse-1-1/): Alibaba's Happy Horse 1.1: a focused model for text, image and reference-driven clips — 3–15 seconds at 720p or 1080p, up to 9 references. Vendor: Alibaba. From $0.168 per second.

## Image models

- [Seedream 5.0 Pro](https://makernice.com/models/seedream-5-0-pro/): ByteDance's Seedream 5.0 Pro: generate or edit images at 1K or 2K, guided by up to 10 reference images. Vendor: ByteDance. From $0.0351 per image.
- [Seedream 5.0](https://makernice.com/models/seedream-5-0/): Blend up to 14 reference images into one: Seedream 5.0 is built for character sheets, collages and style fusion, priced as a single 2K tier. Vendor: ByteDance. From $0.0336 per image.
- [Qwen3 Image](https://makernice.com/models/qwen3-image/): Everyday images and quick edits: Alibaba's Qwen3 Image at 1K or 2K, eight aspect ratios, up to 3 references. Vendor: Alibaba. From $0.0236 per image.
- [Qwen3 Pro Image](https://makernice.com/models/qwen3-pro-image/): Qwen3's premium tier for polished finals — same 1K/2K, aspect ratios and 3 references as Qwen3 Image. Vendor: Alibaba. From $0.0328 per image.

## Blog

- [How to Recreate a Viral AI Video Shot by Shot: A 5-Step Workflow](https://makernice.com/blog/recreate-viral-ai-video/): Recreate a viral short video with AI: split it into shots, make a keyframe per shot, swap in your subject with image-to-image, animate with image-to-video and edit.
- [How to Turn a Reference Video Into Shot-by-Shot AI Video Prompts](https://makernice.com/blog/shot-prompts-from-reference-video/): Break a reference video into a shot list, record five elements per shot, write one prompt per shot and keep each duration inside the model's supported range.
- [Why AI Video Remakes Fail: Fix Faces, Scene Drift and Motion](https://makernice.com/blog/why-ai-remakes-fail/): Why AI video remakes fail and how to fix them: rejected faces, scenes that drift from the original and wrong motion, with the right mode for each fix.
- [Best AI Video Models for Remakes: Key Frames, References, Duration](https://makernice.com/blog/best-models-for-video-remakes/): Which Makernice video models suit a shot-by-shot remake: image-to-video, first and last frame, reference limits, duration, top resolution and audio in one table.
- [Seedance 2.5 API Tutorial: Generate Your First AI Video in 4 Calls](https://makernice.com/blog/seedance-2-5-api-quickstart/): A step-by-step Seedance 2.5 API tutorial: quote the price, create the generation with an idempotency key, poll the status and download the MP4 with curl.
- [How AI Video API Pricing Works: Per Second, Locked by a Quote](https://makernice.com/blog/how-video-pricing-works/): How Makernice prices AI video: unit price times seconds, which fields change the price, how a quote locks it for 12 minutes, and when money is held, charged or released.
- [Text-to-Video Prompt Tips for Seedance 2.5, Wan 3.0 and Kling 3.0](https://makernice.com/blog/text-to-video-prompt-tips/): How to write a text-to-video prompt in five parts (subject, action, camera, light, pace), with weak vs better examples and the aspect ratios each model accepts.
- [Seedance 2.5 vs Kling 3.0: Duration, Resolution, Audio Compared](https://makernice.com/blog/seedance-2-5-vs-kling-3-0/): Seedance 2.5 vs Kling 3.0 on the Makernice API: modes, duration range, resolution, aspect ratios, reference images and audio, side by side in one table.
- [Video Generation API Best Practices: Idempotency, 429s and Polling](https://makernice.com/blog/api-integration-best-practices/): How to integrate an async video generation API safely: idempotency keys, quote expiry, polling with Retry-After, which errors to retry, cancellation and output expiry.
- [Estimate Your AI Video and Image Budget with GET /v1/pricing](https://makernice.com/blog/estimate-your-generation-budget/): Read unit prices from the public GET /v1/pricing endpoint, multiply by seconds or images, and confirm with a quote before you launch a batch on Makernice.
- [Image-to-Video and First-Last Frame API: Fields and Supported Models](https://makernice.com/blog/image-to-video-and-first-last-frame/): How to animate a still with image_url or pin both ends of a clip with last_image_url, which video models support each mode, and which fields to leave out.
- [Wan 3.0 vs MiniMax H3: Duration, Resolution and References](https://makernice.com/blog/wan-3-0-vs-minimax-h3/): Wan 3.0 vs MiniMax H3, plus H3 Max, H3 Max Turbo and Wan 3.0 Prime: duration range, resolution options, aspect ratios and reference-image limits in one table.
- [Reference-to-Video: Image Limits for Seedance 2.5, Wan 3.0 and More](https://makernice.com/blog/reference-to-video-guide/): Which video models support reference-to-video, how many reference images each accepts, and the aspect_ratio rules that cause 400 errors in r2v mode.
- [Seedream 5.0 and Qwen3 Image Pricing: Per Image, What Changes It](https://makernice.com/blog/image-generation-pricing/): How Makernice prices Seedream 5.0, Seedream 5.0 Pro, Qwen3 Image and Qwen3 Pro Image: one price per image, the fields that change it, and how to budget a batch.
- [Seedream 5.0 vs Qwen3 Image: Resolution, Ratios, References](https://makernice.com/blog/seedream-5-0-vs-qwen3-image/): Seedream 5.0 vs Qwen3 Image, plus the Pro variants: resolution handling, aspect ratios and how many reference images image-to-image accepts, in one table.
- [AI Video Duration and Resolution Limits by Model and Their Cost](https://makernice.com/blog/choosing-resolution-and-duration/): The allowed durations and resolutions for Seedance, Wan, MiniMax, Kling and Happy Horse, how cost grows with seconds, and a draft-then-final workflow.

## Links

- [Full Documentation for LLMs](https://makernice.com/llms-full.txt)