25 models live from Kling, Sonilo, BytePlus and MiniMax, one API key

The model API for creative platforms. We supply. We never compete.

Video generation, music, sound effects and dubbing from Kling, Sonilo, BytePlus and MiniMax. One API, one prepaid wallet, billed per second or at the provider’s cost.

Get an API key

Top-ups from $10 · No subscription · Sign up with Google or email

Set up with your coding agent
Read https://scalingtensor.com/SKILL.md and follow it to add ScalingTensor to this project. Ask me for the API key; never hard-code it.
Serving 25 models from 4 providers,
including Kling 3.0 Omni Pro Video Edit, Sonilo Video Dubbing and Dreamina Seedance 2.0

Three steps to your first clip.

  1. Pick a model

    Search, compare prices, try it in the playground
    25Models in the catalog

  2. Call one API

    One key · 4 providers
    1API key for every model

  3. Pay per job

    Prepaid wallet · no subscription
    $10Minimum top-up · 17 clips

How a job runs.

Every model takes the same request and returns the same kind of job. Submit it, poll it, collect the file. The wallet pays for whichever model you picked.

POST/api/media/generate
apiKey: YOUR_API_KEY
{
  "model": Model ID from the catalog"kwaivgi/kling-video/3.0/turbo/standard/text-to-video",
  "params": {
    "prompt": "A red fox trots through fresh snow at dawn, low tracking shot, soft backlight.",
    "duration": 5,
    "aspect_ratio": "16:9"
  }
}
Response · code 0
{ "code": 0,
  "data": { Keep this"jobId": "724958320123",
            "status": "pending" },
  "msg": "Success" }
Tracking one job724958320123 · Kling 3.0 Turbo Standard Text to VideoSubmitted · code 0
Tracking one job724958320123 · Kling 3.0 Turbo Standard Text to VideoSubmitted · code 0

What juggling four providers actually costs you.

Four accounts, four balances, four ways of saying a job is done.

Four wallets run dry on four different days.

Each provider bills you on its own. You top up four balances and learn which one hit zero when a job fails.

Four ways to sign a request.

Bearer tokens, signed JWTs, key pairs. Every provider wants its own.

Four ways to say a job is done.

One says succeed, one says SUCCESS, one returns a number. Your poller has to know them all.

Every model, one catalog.

25 models from four providers across 14 task types. Each one has a page with its price, a request example and a playground.

Browse all 25 models

Running real volume? Apply for partner pricing ›

Jobs run around the clock. So does the API.

Submit from any timezone. Jobs queue with the provider and the result waits for your next poll.

models live
25
providers
4
task types
14
SubmittedPendingCompletedExample jobs for illustration. The band marks local noon as the day turns.

Can't we sign up with each provider? Sure, if you like four invoices.

Otherwise, one prepaid wallet pays for every model, at the price shown on its model page.

Listed prices · USDAll 25 prices ›
WalletOne balance, every model
$4.31

Top-ups from $10 by card through Stripe. No subscription.

  • Top-upcard through Stripe+$10.00
  • Kling 3.0 Omni Pro Video Edit5 s clip · 16:9 · reference price≈ −$0.84
  • Sonilo Video Dubbing20 s × 2 languages · $0.0985/s−$3.94
  • Kling 3.0 Omni Pro Video Reference5 s clip · 16:9 · reference price≈ −$0.84
  • Sonilo Text to Music30 s track · $0.00225/s−$0.0675

Example usage at listed prices. ≈ marks a reference price: those models are billed at the provider’s actual cost.

Sonilo Text to Sound Effects, 30-second output: 30 s × $0.0018 = $0.054

Sonilo Text to Music, 30-second output: 30 s × $0.00225 = $0.0675

Questions developers ask before they integrate.

What is ScalingTensor?

ScalingTensor is an API marketplace for AI models. You call every model in the catalog through one HTTP/JSON API with a single API key, and pay per request from a prepaid wallet.

Where can I find available models and API pricing?

The model catalog lists every enabled model with its configured price per job, per second, per token or at the provider’s actual cost. Each model page also has a request example and a playground to try it before you integrate.

Can I use the OpenAI SDK?

Not directly. Media models use POST /api/media/generate and GET /api/media/result, and the chat endpoint is POST /api/chat/completion. All of them take an apiKey header and return code and data rather than the OpenAI SDK format, so an adapter is required.

How does pricing work?

Each model page shows how it is billed: per job, per second of audio or video, per output quantity, per token, or at the upstream provider’s actual cost. Charges are deducted from your prepaid wallet.

How do I create an account?

Sign up with Google or your email address on the sign-up page; email sign-ups confirm the address with a one-time code. You can then create API keys and top up your wallet.

Is my data secure?

All API traffic uses HTTPS, and requests are authenticated with API keys that you can revoke at any time in the console. Keep keys on your server, never in browser code. Generation requests are processed by the model’s provider, so review that provider’s data policy before sending sensitive content.

Pick a model. See what it costs.

Search a model to see its price, a request example and a playground.

Get an API key

Top-ups from $10 · No subscription · Sign up with Google or email