Pricing

Pay only for what you generate

No subscriptions and no minimums. Every model publishes its unit price before you call it, and each response reports the exact amount settled against your balance.

Units

Billing follows the natural unit of each modality

You are never charged a flat rate for a model you barely use. Video bills by the second, images by the call, and LLMs by the token.

ModalityBilling unitExample modelFrom
Videoper secondVeo 3.1 Fast$0.06
Imageper callGPT Image 2$0.03
Musicper songSuno v5.5$0.18
Audioper 1K charactersElevenLabs TTS v3$0.04
LLMper 1M input tokensGPT-5.6$2.50
Embeddingsper 1M tokensEmbedding 4 Large$0.13

Included

Everything in the base rate

The platform features are not a paid tier. Keys, callbacks, and analytics ship with every account.

  • Unlimited API keys, each with its own budget and scopes
  • Async task management with automatic retries and refunds
  • Webhook callbacks with signed deliveries
  • Usage analytics and per-key cost tracking

Units

Everything in the base rate

The platform features are not a paid tier. Keys, callbacks, and analytics ship with every account.

Failed generations are free

If a provider fails or filters a request, the reserved credit is released. You are only billed for delivered output.

No token markup games

Input and output prices are published per model, so you can compute the cost of a request before sending it.

Budgets as guardrails

Set a monthly cap per key. When it is reached, requests return 402 instead of silently spending.

Volume terms for teams

Team and enterprise plans add custom rate limits, invoicing, and a contractual uptime SLA.

Building something large?

High-volume commitments qualify for reduced unit pricing and dedicated throughput. Tell us the shape of your workload.

Contact sales

Start with free credits

Create an account, generate a key, and see exactly what each call costs before you commit.