# AirCube

> AirCube is a pay-as-you-go API, operated by Vansor Technology Limited in Hong Kong, that runs image, video and audio generation models from many providers, such as Seedream, Kling, Veo, Sora and ElevenLabs, behind one key and one bill. Requests are plain REST calls that you submit and then poll, with no SDK needed. New accounts get free credits with no card, and each generation is billed at the model's listed rate with no subscription.

Page: https://scratchdb.com/tools/aircube/
Facts checked on 20 November 2026.

## Key facts
- Website: https://aircube.ai
- Made by: Vansor Technology Limited
- Category: AI Platforms & APIs
- Pricing: Pay as you go
- Free plan: Free credits for new accounts, no card needed. The amount isn't published.
- Paid from: Per generation, for example $0.03 an image or $0.20 a second of video, depending on the model
- Open source: no

## A good fit for
- Developers adding image or video generation who want to try several models without separate vendor accounts
- Apps that switch between models as new ones launch and want the code to stay the same
- Small teams that prefer one invoice and per-generation billing to monthly plans

## Not a good fit for
- Text chat or coding models, since the API focuses on image, video and audio generation
- Teams that want the free credit amount or volume prices up front, since the site publishes neither
- Projects that need an official SDK, since the API is plain REST only

## Pricing
Pay as you go: each generation is billed at the model's listed rate, with no subscription or minimum. New accounts get free credits without a card, and high-volume customers can get custom rates and invoicing.

| Plan | Price | What you get |
|---|---|---|
| seedream-v5.0-pro (image) | $0.045 an image | Text-to-image. |
| qwen-image-2.1 (image) | $0.03 an image | Text-to-image. |
| gpt-image-2 edit (image) | $0.0665 an image | Image editing. |
| wan-3.0 (video) | $0.20 a second | Image-to-video. |
| seedance-2.5 (video) | $0.76 a second | Image-to-video. |
| q3 (video) | $0.43 a second | Image-to-video. |
| Enterprise | Contact sales | Custom rates, higher limits, dedicated support and invoicing. |

Official prices: https://aircube.ai/pricing

## Main features
- One API for many models: Image, video and audio models from several providers behind one key and one consistent request format.
- Plain REST: Submit with a POST to the model's endpoint and poll /api/v3/status/{id}; no SDK required.
- Per-model pricing: Each model has a listed price per image or per second, for example $0.045 an image on seedream-v5.0-pro.
- Video generation: Text-to-video and image-to-video models such as Kling, Seedance, Veo, Sora and Wan.
- Audio and face swap: Speech and music with ElevenLabs, Qwen3 TTS and MiniMax Music, plus a detect-then-swap face swap workflow.
- No-code generator: A browser tool that runs the same models without writing code.

## About
AirCube puts generation models from many labs behind a single REST API. You send a request to a model's endpoint with your AirCube key, get back a job id, and poll a status endpoint until the image, video or audio is ready. Switching models means changing the endpoint, not adding a new vendor SDK or contract.

The docs cover image generation and editing (Seedream, Flux, GPT Image and others), video from text or images (Kling, Seedance, Veo, Sora), speech and music (ElevenLabs, Qwen3 TTS, MiniMax Music) and face swap. The homepage says 450+ models, while the documentation says 87+ models are available through the API.

Billing is per image, per second of video or per token at each model's listed rate, on one invoice. There is also a no-code AI Generator in the browser that uses the same models, for people who don't write code.

## Alternatives
- [Anthropic](https://scratchdb.com/tools/anthropic/): AI research company that makes the Claude models, the Claude apps and the Claude API.
- [OpenAI](https://scratchdb.com/tools/openai/): AI research company behind ChatGPT and the GPT model family, accessed through apps and an API.
- [OpenAI API](https://scratchdb.com/tools/openai-api/): Pay-per-token API access to OpenAI's GPT models for building AI into your own applications.
- [Ollama](https://scratchdb.com/tools/ollama/): Run open source AI models locally, with an optional hosted cloud and Pro plan.
- [Hugging Face](https://scratchdb.com/tools/huggingface/): Hub for open source AI models, datasets and demo apps, plus paid hosting for inference and Spaces.
- [Mistral AI](https://scratchdb.com/tools/mistral-ai/): French AI company building open-weight and proprietary language models, accessed via API or Le Chat.
- [Groq](https://scratchdb.com/tools/groq/): AI inference company behind GroqCloud, a fast API for running open models on its own LPU chips.
- [Cohere](https://scratchdb.com/tools/cohere/): Enterprise-focused AI company offering language, embedding and rerank models through an API.

All alternatives side by side: https://scratchdb.com/alternatives/aircube/

## Questions
### Is AirCube free?
New accounts get free credits with no card required. After that, each generation is billed at the model's listed rate.

### How does AirCube pricing work?
It is pay as you go, per image, per second of video or per token, with no subscription or minimum spend. Prices range from $0.03 an image to $0.76 a second of video on the models shown on its pricing page.

### Which models does AirCube support?
Image models such as Seedream, Flux and GPT Image, video models such as Kling, Seedance, Veo and Sora, and audio models such as ElevenLabs. Its docs say 87+ models are available through the API.

### Does AirCube have an SDK?
No SDK is needed. It is a REST API you call with a bearer key from any language.

## Sources
- [AirCube](https://aircube.ai/)
- [AirCube pricing](https://aircube.ai/pricing)
- [AirCube documentation](https://aircube.ai/docs)
- [AirCube terms and conditions](https://aircube.ai/legal/terms)
