Thinking Frog 中文ENFR
Claude

One key, call Claude.

0.7
stream

Thinking Frog is an AI API hub: one OpenAI-compatible endpoint for Claude, GPT, Gemini, DeepSeek and the other major models. All three settings above are live.

Models

Seven providers, one endpoint, one key. Tap a name and it goes into the sentence at the top of the page.

From application to first call

  1. About two minutes

    ApplyFirst, fill in the form at the bottom of this page.

  2. One message

    Contact usThen message customer service on WeChat or by email with your name.

  3. Usually within one working day

    ActivationNext, we confirm your use case, models and quota, and issue your API key.

  4. One line of code

    First callFinally, point base_url at our endpoint and start calling.

What you get

OpenAI-compatible

Requests and responses match the OpenAI format, so the official SDKs and LangChain work as they are.

One key for all

No separate sign-ups, top-ups or keys for each provider: one account covers every model in the catalogue.

Streaming

Turn on stream and answers arrive as they are generated, just like the switch at the top.

Tool calling

Function calling is written exactly as with OpenAI; what each model supports depends on the model.

Usage records

Every call is logged with its model, tokens and cost, so the monthly bill holds no surprises.

Personal support

Accounts are opened by a person; quota, models and billing are agreed around your business.

FAQ

Why contact customer service after applying? Accounts are activated by hand. Once we have confirmed your use case, the models you need and your expected volume, we issue your API key. This prevents abuse and lets us set the right quota and billing for you.

Do I need to rewrite my code? No. The endpoint is OpenAI-compatible: usually you only change base_url to the Thinking Frog address and api_key to the key you receive.

Which models are available? The catalogue lists the providers and common models we connect to today; the list in your dashboard after activation is the reference. If a model you need is missing, say so in the notes of your application.

How is usage billed? You pay for what you use, and prices differ by model. Customer service will give you prices and billing options when your account is opened.

Are streaming and function calling supported? Yes, with the same parameters as OpenAI. A feature a model does not support itself is not available through us either.

Are there rate or quota limits? Each account runs on the quota agreed at activation. When your volume grows, contact customer service to raise it.

Can one account serve several projects? Yes. Tell customer service about your projects when you apply and they will help you plan keys and quotas.

Who is the contracting and invoicing party? Thinking Frog is operated by Thinking Frog Limited. Contracts, billing and invoices are confirmed with customer service when your account is opened.

Apply for access

Fill in the lines below and submit, then contact customer service with your name. We will open your account and send your API key.

Models you want

We only use these details to contact you about your account.