> ## Documentation Index
> Fetch the complete documentation index at: https://docs.anyone.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# What are tokens

> Understand the difference between API keys and model tokens in Anyone.

"Token" has two meanings in Anyone — don't mix them up:

## API key (your credential)

An API key is the credential you use to call the Anyone API. It typically looks like `sk-xxxxx`.

* Created in the dashboard
* Included in the `Authorization: Bearer` header of your requests
* Each API key has its own usage tracking and quota
* You can create new ones and delete old ones at any time

See [Manage your API keys](/guides/manage-tokens) for details.

## Model tokens (billing unit)

Model tokens are the basic units AI models use to process text. Roughly:

* 1 English word ≈ 1–2 tokens
* 1 Chinese character ≈ 1–3 tokens
* 1,000 tokens ≈ 750 English words

The cost of each API call = input tokens × input price + output tokens × output price

The `usage` field in the API response tells you how many tokens each request consumed:

```json theme={null}
{
  "usage": {
    "prompt_tokens": 25,      // 25 tokens for input
    "completion_tokens": 100,  // 100 tokens for output
    "total_tokens": 125        // 125 tokens total
  }
}
```

<Tip>
  Want to save money? Use shorter system prompts, reduce context length, and choose cheaper models (e.g., gpt-5.4-mini instead of gpt-5.4).
</Tip>