SeaWhale AI

English

简体中文

English

简体中文

Appearance

Sign in

FAQ ​

Getting started ​

How is SeaWhale AI different from other AI API gateways?

Three main differences:

  • Choice of channels — most models are available over direct, preferred and economy channels, and you pick based on cost and reliability rather than having the platform decide for you. Prices for the same model can differ several-fold between channels, so routing batch jobs through the economy channel meaningfully lowers costs.
  • Transparent pricing — per-model prices are published, billing is based on actual token usage, and there are no monthly fees, mandatory plans or expiring balances.
  • Reliable — we aggregate several upstream providers and fall back automatically when a link fails. The API is fully OpenAI SDK compatible, so migrating usually means changing just base_url and api_key.

See Quick start for how to use channels.

Do I need my own overseas account to use the gateway?

No. A SeaWhale AI key gives you access to every model. You can top up locally without an overseas account or credit card, and you do not need to integrate separately with OpenAI, Anthropic and Google authentication and protocols.

Why choose SeaWhale AI?

SeaWhale AI gives you one API that reaches every mainstream large language model on the market.

Key advantages:

  • One integration — call every major AI model through a single API
  • Unified billing — costs for all models in one place, easy to manage
  • Usage analytics — track and analyze all API calls in real time
  • Transparent pricing — based on the upstream provider's own pricing, with no markup
  • High availability — aggregated provider capacity with automatic failover keeps the service stable
How do I get started with SeaWhale AI?

Step 1: create an account

  • Visit the SeaWhale AI website and sign up

Step 2: add credit

  • Add credit on the billing page (credit is a prepayment for model usage on the platform)
  • Each API call deducts credit based on the number of tokens used
  • Pricing varies by model — see the model list for details

Step 3: start building

  • Create an API key and follow the quick start guide to integrate it into your application
How do I get technical support?

Technical questions:

  • Check our online documentation and API reference

Billing and account questions:

Contact support:

  • If you run into trouble, scan the QR code to reach our support team on WeChat Work and an engineer will help you complete the integration
Support

Pricing and billing ​

How does SeaWhale AI bill?

Pricing model:

  • Every model is priced per thousand tokens (1k tokens)
  • Input tokens (prompt) and output tokens (completion) are usually priced differently
  • Some models bill separately for images, reasoning tokens and similar
  • See the model list page for detailed pricing

Billing flow:

  1. You send an API request
  2. The provider processes it and reports the tokens actually used
  3. The system calculates the cost and deducts it from your credit balance
  4. You can review your full usage history and cost breakdown in the account center

Transparent billing: every request shows detailed token usage and cost components, so you can see exactly where the money went.

Models ​

Which AI models does SeaWhale AI support?

SeaWhale AI provides access to nearly every mainstream AI model, including:

Major providers:

  • OpenAI (GPT-5, GPT-4 and more)
  • Anthropic (the Claude family)
  • Google (the Gemini family)
  • Meta (the Llama family)
  • Plus frontier models from other leading AI labs

How to browse:

  • Visit the model browser for the full list and detailed information
  • Fetch the available model list programmatically through the models API
  • Each model lists its price, context length and supported capabilities
How quickly are new models added?

Update cadence:

  • We work closely with the major AI labs
  • New models are added as soon as possible after release, usually within 1–2 days
  • Significant models are prioritized

Model requests:

  • We schedule new integrations based on user demand

Continuous improvement: we do not only add new models — we keep tuning the performance and stability of existing ones.

API specifications ​

Which authentication methods are supported?

SeaWhale AI supports the following:

1. Cookie authentication

  • Used by the web interface and online chat
  • Managed automatically after you sign in

2. API key authentication

  • Used for API calls
  • Passed as a token in the HTTP request headers
  • Format: Authorization: YOUR_API_KEY

Security note: keep your API key safe and never expose it in public code.

Are there rate limits?

Model limits:

  • To keep model access fair across users, SeaWhale AI applies baseline rate limiting. Limits are per model and tied to your main SeaWhale AI account, counting the combined calls to that model from every API key under the account. Requests beyond the limit fail and must be retried once the limit resets.

When you exceed a limit:

  • You receive a 429 error code
  • We recommend implementing retries and client-side rate control
Which input formats does the API support?

Supported today:

  • Text: plain text input in any language
  • Images: URLs or base64-encoded image data

Coming soon:

  • PDF documents
  • Audio files
  • Video files
  • Other common document formats

Image formats: JPG, PNG, WebP and other common formats are supported, though exact support depends on the model you choose.