### Quickstart Installation

Source: https://ezaiapi.com/docs

Follow these steps to quickly set up EzAI and start using it with Claude Code.

```APIDOC
## Quickstart
Up and running in under a minute. Sign up to get your API key.

1.  ### Get your API key
    Sign in and copy your API key from the dashboard.

2.  ### Run the install command
    Sets `ANTHROPIC_BASE_URL`, `ANTHROPIC_API_KEY`, and configures `~/.claude/settings.json`.

    **macOS / Linux**
    ```bash
    curl -fsSL "/install.sh?key=YOUR_KEY" | sh
    ```

    **Windows**
    ```powershell
    irm "/install.ps1?key=YOUR_KEY" | iex
    ```

    View source: install.sh · install.ps1

3.  ### Restart your terminal
    ```bash
    source ~/.bashrc  # or ~/.zshrc
    ```

4.  ### Start using Claude Code
    ```bash
    claude
    ```

    That's it! Claude Code routes through EzAI automatically.
```

--------------------------------

### Start Claude Code

Source: https://ezaiapi.com/docs

Once configured, you can start using Claude Code, which will automatically route through EzAI.

```bash
claude
```

--------------------------------

### Install EzAI CLI (Windows PowerShell)

Source: https://ezaiapi.com/docs

Installs the EzAI CLI tool on Windows using PowerShell.

```powershell
irm "/install.ps1?key=YOUR_KEY" | iex
```

--------------------------------

### Manual Installation

Source: https://ezaiapi.com/docs

Alternatively, you can manually set up EzAI by configuring environment variables and settings files.

```APIDOC
## Manual Installation
Prefer manual setup? Two steps:

### 1. Set environment variables
Add to `~/.bashrc`, `~/.zshrc`, or your shell config:
```bash
export ANTHROPIC_BASE_URL="https://ezaiapi.com"
export ANTHROPIC_API_KEY="YOUR_KEY"
```

### 2. Update settings.json
Create or edit `~/.claude/settings.json`:
```json
{
  "env": {
    "ANTHROPIC_BASE_URL": "https://ezaiapi.com",
    "ANTHROPIC_API_KEY": "YOUR_KEY"
  },
  "disableLoginPrompt": true
}
```
```

--------------------------------

### Install EzAI CLI (Windows)

Source: https://ezaiapi.com/docs

Installs the EzAI CLI tool on Windows systems using PowerShell.

```powershell
curl -fsSL "https://ezaiapi.com/install.sh?key=YOUR_KEY" | sh
```

--------------------------------

### Install EzAI CLI (macOS/Linux)

Source: https://ezaiapi.com/docs

Installs the EzAI CLI tool, setting necessary environment variables and configuration for Claude Code integration.

```bash
curl -fsSL "/install.sh?key=YOUR_KEY" | sh
```

--------------------------------

### Restart Terminal

Source: https://ezaiapi.com/docs

After installation, restart your terminal or source your shell configuration file to apply changes.

```bash
source ~/.bashrc  # or ~/.zshrc
```

--------------------------------

### Compatibility with Tools

Source: https://ezaiapi.com/docs

EzAI is designed to be a drop-in replacement for Anthropic and OpenAI-compatible tools. This section details the setup for various popular tools.

```APIDOC
## Compatibility
EzAI works as a drop-in replacement for any Anthropic or OpenAI-compatible tool.

Tool | Status | Setup  
---|---|---
Claude Code| ✅ Native| Set env vars or run install script  
Cursor| ✅ Works| Settings → Models → Custom API provider  
Cline| ✅ Works| Extension settings → Anthropic → Base URL  
Continue| ✅ Works| config.json → custom provider  
Aider| ✅ Works| Set ANTHROPIC_BASE_URL env var  
OpenAI SDK| ✅ Works| Set custom User-Agent (see above)  
LiteLLM| ✅ Works| Set api_base + api_key  
Any Anthropic tool| ✅ Works| Change base URL + API key  
```

--------------------------------

### OpenAI SDK Integration

Source: https://ezaiapi.com/docs

Demonstrates how to integrate with the EzAI API using the official OpenAI SDK, including necessary configurations for custom User-Agent.

```APIDOC
## OpenAI SDK Integration

Use the official OpenAI SDK to access all EzAI models via the `/v1/chat/completions` endpoint.

**⚠️ Important:** Cloudflare blocks the default `User-Agent: OpenAI/Python` header with a 403 error. You **must** set a custom User-Agent as shown below.

### Python
```python
# pip install openai
from openai import OpenAI

client = OpenAI(
    base_url="https://ezaiapi.com/v1",
    api_key="sk-your-key",
    default_headers={"User-Agent": "EzAI/1.0"}  # Required!
)

response = client.chat.completions.create(
    model="claude-sonnet-4-5",
    messages=[
        {"role": "system", "content": "You are helpful."},
        {"role": "user", "content": "Hello!"}
    ],
    max_tokens=1024
)
print(response.choices[0].message.content)
```

### Node.js
```javascript
// npm install openai
import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://ezaiapi.com/v1',
  apiKey: 'sk-your-key',
  defaultHeaders: { 'User-Agent': 'EzAI/1.0' }  // Required!
});

const resp = await client.chat.completions.create({
  model: 'claude-sonnet-4-5',
  messages: [
    { role: 'system', content: 'You are helpful.' },
    { role: 'user', content: 'Hello!' }
  ],
  max_tokens: 1024
});
console.log(resp.choices[0].message.content);
```
```

--------------------------------

### OpenAI SDK Integration

Source: https://ezaiapi.com/docs

Use the official OpenAI SDK. Note: A custom User-Agent header is required to bypass Cloudflare restrictions.

```python
# pip install openai
from openai import OpenAI

client = OpenAI(
    base_url="https://ezaiapi.com/v1",
    api_key="sk-your-key",
    default_headers={"User-Agent": "EzAI/1.0"}  # Required!
)

response = client.chat.completions.create(
    model="claude-sonnet-4-5",
    messages=[
        {"role": "system", "content": "You are helpful."},
        {"role": "user", "content": "Hello!"}
    ],
    max_tokens=1024
)
print(response.choices[0].message.content)
```

```javascript
// npm install openai
import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://ezaiapi.com/v1',
  apiKey: 'sk-your-key',
  defaultHeaders: { 'User-Agent': 'EzAI/1.0' }  // Required!
});

const resp = await client.chat.completions.create({
  model: 'claude-sonnet-4-5',
  messages: [
    { role: 'system', content: 'You are helpful.' },
    { role: 'user', content: 'Hello!' }
  ],
  max_tokens: 1024
});
console.log(resp.choices[0].message.content);
```

--------------------------------

### Configure OpenAI SDK with Custom User-Agent

Source: https://ezaiapi.com/docs

Set a custom User-Agent to bypass Cloudflare WAF blocks that target the default OpenAI Python header.

```python
client = OpenAI(
    base_url="https://ezaiapi.com/v1",
    api_key="sk-your-key",
    default_headers={"User-Agent": "EzAI/1.0"}
)
```

--------------------------------

### Chat Completions via OpenAI Format

Source: https://ezaiapi.com/docs

Interact with models using the standard OpenAI Chat Completions API format.

```bash
curl https://ezaiapi.com/v1/chat/completions \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "messages": [
      { "role": "system", "content": "You are helpful." },
      { "role": "user", "content": "Hello!" }
    ],
    "max_tokens": 1024
  }'
```

```json
{
  "id": "req_abc123",
  "object": "chat.completion",
  "model": "claude-sonnet-4-5",
  "choices": [{
    "index": 0,
    "message": { "role": "assistant", "content": "Hello! How can I help?" },
    "finish_reason": "stop"
  }],
  "usage": { "prompt_tokens": 15, "completion_tokens": 9, "total_tokens": 24 }
}
```

--------------------------------

### Update settings.json (Manual)

Source: https://ezaiapi.com/docs

Configure Claude Code settings by creating or editing the `~/.claude/settings.json` file to include EzAI credentials.

```json
{
  "env": {
    "ANTHROPIC_BASE_URL": "https://ezaiapi.com",
    "ANTHROPIC_API_KEY": "YOUR_KEY"
  },
  "disableLoginPrompt": true
}
```

--------------------------------

### Troubleshooting Common Issues

Source: https://ezaiapi.com/docs

Provides solutions for specific troubleshooting scenarios, including 403 errors with the OpenAI SDK, IP blocks, rate limit exceedances, connection timeouts, and Claude Code configuration.

```APIDOC
## Troubleshooting
### 403 Forbidden with OpenAI SDK
Cloudflare WAF blocks the default `User-Agent: OpenAI/Python` header. Set a custom User-Agent:
```python
client = OpenAI(
    base_url="https://ezaiapi.com/v1",
    api_key="sk-your-key",
    default_headers={"User-Agent": "EzAI/1.0"}
)
```

### 403 Forbidden — IP Blocked
Your IP was auto-blocked after 50+ failed auth attempts. **Wait 5 minutes** for the block to expire, then fix your API key. Check the key in your dashboard.
### 429 Too Many Requests
You've exceeded your tier's rate limit (RPM or concurrent). Check the `Retry-After` header for when to retry. If your balance is exhausted (402), top up in the dashboard. See Rate Limits for your tier's limits.
### Connection Timeout / Reset
Intermittent upstream connection issues. The proxy automatically retries. If persistent, try again after a few seconds — the upstream model may be under heavy load.
### Claude Code not using proxy
Verify your environment variables are set:
```bash
echo $ANTHROPIC_BASE_URL   # Should show https://ezaiapi.com
echo $ANTHROPIC_API_KEY  # Should show your API key
cat ~/.claude/settings.json # Should contain env block
```
```

--------------------------------

### Set Environment Variables (Manual)

Source: https://ezaiapi.com/docs

Manually set the ANTHROPIC_BASE_URL and ANTHROPIC_API_KEY environment variables in your shell configuration.

```bash
export ANTHROPIC_BASE_URL="https://ezaiapi.com"
export ANTHROPIC_API_KEY="YOUR_KEY"
```

--------------------------------

### Verify Claude Code Proxy Configuration

Source: https://ezaiapi.com/docs

Check environment variables and settings to ensure Claude Code is correctly routing through the EzAI proxy.

```bash
echo $ANTHROPIC_BASE_URL   # Should show https://ezaiapi.com
echo $ANTHROPIC_API_KEY  # Should show your API key
cat ~/.claude/settings.json # Should contain env block
```

--------------------------------

### OpenAI Compatible Chat Completions

Source: https://ezaiapi.com/docs

Provides an OpenAI-compatible endpoint for chat completions, allowing integration with OpenAI SDKs and tools like LiteLLM.

```APIDOC
## POST /v1/chat/completions

### Description
Native OpenAI Chat Completions format. Use with OpenAI SDK, LiteLLM, or any OpenAI-compatible tool.

### Method
POST

### Endpoint
`/v1/chat/completions`

### Parameters
#### Request Body
- **model** (string) - Required - The model to use for the chat completion.
- **messages** (array) - Required - An array of message objects.
  - **role** (string) - Required - The role of the message sender (e.g., "system", "user", "assistant").
  - **content** (string) - Required - The content of the message.
- **max_tokens** (integer) - Optional - The maximum number of tokens to generate in the completion.

### Request Example
```json
{
  "model": "claude-sonnet-4-5",
  "messages": [
    { "role": "system", "content": "You are helpful." },
    { "role": "user", "content": "Hello!" }
  ],
  "max_tokens": 1024
}
```

### Response
#### Success Response (200)
- **id** (string) - Unique identifier for the completion.
- **object** (string) - Type of object returned (e.g., "chat.completion").
- **model** (string) - The model used for the completion.
- **choices** (array) - An array of completion choices.
  - **index** (integer) - Index of the choice.
  - **message** (object) - The message content and role.
    - **role** (string) - The role of the message sender (e.g., "assistant").
    - **content** (string) - The content of the assistant's message.
  - **finish_reason** (string) - The reason the generation finished (e.g., "stop").
- **usage** (object) - Token usage information.
  - **prompt_tokens** (integer) - Number of tokens in the prompt.
  - **completion_tokens** (integer) - Number of tokens in the completion.
  - **total_tokens** (integer) - Total number of tokens used.

#### Response Example
```json
{
  "id": "req_abc123",
  "object": "chat.completion",
  "model": "claude-sonnet-4-5",
  "choices": [
    {
      "index": 0,
      "message": { "role": "assistant", "content": "Hello! How can I help?" },
      "finish_reason": "stop"
    }
  ],
  "usage": { "prompt_tokens": 15, "completion_tokens": 9, "total_tokens": 24 }
}
```
```

--------------------------------

### Streaming Responses

Source: https://ezaiapi.com/docs

Enable streaming responses by setting `"stream": true`. This returns Server-Sent Events for real-time updates.

```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "max_tokens": 1024,
    "stream": true,
    "messages": [
      { "role": "user", "content": "Write a haiku" }
    ]
  }'
```

--------------------------------

### List Available Models

Source: https://ezaiapi.com/docs

Retrieves a list of all available models in OpenAI format, useful for model discovery.

```APIDOC
## GET /v1/models

### Description
Returns all available models in OpenAI format. Useful for model discovery in tools like LiteLLM.

### Method
GET

### Endpoint
`/v1/models`

### Alternate Endpoints
- `/openai/v1/models`
- `/api/v1/models`

### Single Model Endpoint
- **GET /v1/models/{model_id}** - Returns model info or 404.
```

--------------------------------

### Rate Limits and Response Headers

Source: https://ezaiapi.com/docs

Understand the rate limits applied per account based on your tier, including RPM, concurrent requests, and daily limits. Response headers provide real-time usage information.

```APIDOC
## Rate Limits
Rate limits are applied **per account** based on your tier. Two limits apply simultaneously:

RPM
Requests per minute
Concurrent
Max parallel requests
5 min
IP block auto-expires

Tier | RPM | Concurrent | Daily Requests | Condition  
---|---|---|---|---
Free| 5| 1| 50| $0 balance, no plan  
Top-up| 30| 2| 500| Has credit balance  
Starter| 30| 2| 1,000| $10/month plan  
Pro| 60| 3| 2,000| $20/month plan  
Max| 90| 6| 5,000| $40/month plan  
Ultra| 120| 8| 10,000| $80/month plan  

**💎 Monthly Plans** — Get free credits that reset every 5 hours + higher rate limits + daily request quota. Multiple plans stack daily limits. View plans →

### Response Headers
Every API response includes rate limit headers so your client can track usage:
```
x-ratelimit-tier: topup
x-ratelimit-limit-requests: 30          # RPM limit for your tier
x-ratelimit-remaining-requests: 28      # Remaining requests this minute
x-ratelimit-reset-requests: 2026-03-03T09:00:00.000Z
x-concurrent-limit: 5                   # Max parallel requests
x-concurrent-remaining: 3               # Available parallel slots
```

* •Exceeding any limit returns `429` with a `Retry-After` header
* •50+ consecutive failed auth attempts from the same IP triggers auto-block (`403`)
* •IP blocks expire automatically after 5 minutes
* •Balance exhaustion returns `402` — top up
* •Your current tier is shown in your dashboard — upgrade plan for higher limits
```

--------------------------------

### Create Message (GPT)

Source: https://ezaiapi.com/docs

Send a message to the GPT-4.1 model using the EzAI API. The proxy auto-converts to OpenAI format upstream.

```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "gpt-4.1",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Hello, GPT!" }
    ]
  }'
```

--------------------------------

### Responses API Integration

Source: https://ezaiapi.com/docs

Use the simplified Responses API for compatibility with tools like n8n and LangChain.

```bash
curl https://ezaiapi.com/v1/responses \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{ "model": "claude-sonnet-4-5", "input": "Hello!", "max_output_tokens": 1024 }'
```

--------------------------------

### Enable Extended Thinking (Chain-of-Thought)

Source: https://ezaiapi.com/docs

Activate Claude's chain-of-thought reasoning for complex tasks by including the `thinking` object in the request. Recommended with streaming and a budget of 5K-10K tokens.

```json
{
  "model": "claude-opus-4-6",
  "max_tokens": 16000,
  "stream": true,
  "thinking": {
    "type": "enabled",
    "budget_tokens": 10000
  },
  "messages": [
    { "role": "user", "content": "Solve this step by step..." }
  ]
}
```

--------------------------------

### Supported Models

Source: https://ezaiapi.com/docs

Lists the AI models supported by the EzAI API, categorized by provider.

```APIDOC
## Supported Models

All models are accessible via `/v1/messages` (Anthropic), `/v1/chat/completions` (OpenAI) and `/v1/responses` (Responses API) formats.

### Claude
- claude-opus-4-6 (144K context)
- claude-opus-4-5 (160K context)
- claude-sonnet-4-6 (new, 200K context)
- claude-sonnet-4-5 (144K context)
- claude-sonnet-4 (216K context)
- claude-haiku-4-5 (144K context)

### GPT / OpenAI
- gpt-5.4 (new, ★, 400K context)
- gpt-5.3-codex (400K context)
- gpt-5.2 (264K context) / gpt-5.2-codex (400K context)
- gpt-5.1 (264K context) / gpt-5.1-codex (400K context)
- gpt-5.1-codex-mini / max (400K context)
- gpt-5 / gpt-5-mini (264K context)
- gpt-4.1 / gpt-4.1-mini / nano (128K context)
- gpt-4o / gpt-4o-mini (128K context)
- o3 / o3-mini / o4-mini (aliases)

### Gemini
- gemini-3.1-pro (preview, 128K context)
- gemini-3-pro (preview, 128K context)
- gemini-3-flash (preview, 128K context)
- gemini-2.5-pro (128K context)

### xAI
- grok-code-fast-1 (128K context)
```

--------------------------------

### Streaming Responses

Source: https://ezaiapi.com/docs

Enable streaming to receive responses as Server-Sent Events (SSE), which is useful for real-time interactions.

```APIDOC
## Streaming
Add `"stream": true` to receive Server-Sent Events:

```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{ 
    "model": "claude-sonnet-4-5",
    "max_tokens": 1024,
    "stream": true,
    "messages": [
      { "role": "user", "content": "Write a haiku" }
    ]
  }'
```

SSE events: `message_start` → `content_block_delta` → `message_stop`. See Anthropic streaming docs.
```

--------------------------------

### Anthropic SDK Integration

Source: https://ezaiapi.com/docs

Access multiple model families using the official Anthropic SDK by pointing the base_url to EzAI.

```python
# pip install anthropic
import anthropic

client = anthropic.Anthropic(
    api_key="sk-your-key",
    base_url="https://ezaiapi.com"
)

# Claude
msg = client.messages.create(
    model="claude-sonnet-4-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}]
)
print(msg.content[0].text)

# GPT — same client, just change model
msg = client.messages.create(model="gpt-4.1", max_tokens=1024,
    messages=[{"role": "user", "content": "Hello GPT!"}])

# Gemini
msg = client.messages.create(model="gemini-2.5-pro", max_tokens=1024,
    messages=[{"role": "user", "content": "Hello Gemini!"}])
```

```javascript
// npm install @anthropic-ai/sdk
import Anthropic from '@anthropic-ai/sdk';

const client = new Anthropic({
  apiKey: 'sk-your-key',
  baseURL: 'https://ezaiapi.com'
});

const msg = await client.messages.create({
  model: 'claude-sonnet-4-5',
  max_tokens: 1024,
  messages: [{ role: 'user', content: 'Hello!' }]
});
console.log(msg.content[0].text);
```

--------------------------------

### Count Tokens via API

Source: https://ezaiapi.com/docs

Calculate the number of tokens for a given message payload without incurring credit charges.

```bash
curl https://ezaiapi.com/v1/messages/count_tokens \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{ "model": "claude-sonnet-4-5", "messages": [{ "role": "user", "content": "Hello!" }] }'
```

--------------------------------

### Error Codes and Handling

Source: https://ezaiapi.com/docs

This section details common error codes returned by the API, their meanings, and recommended actions to resolve them.

```APIDOC
## Error Codes
Code | Meaning | What to Do  
---|---|---
400| Bad request / invalid body| Check your JSON payload and required fields  
401| Invalid or missing API key| Check `x-api-key` header value  
402| Insufficient balance| Top up credits in dashboard  
403| IP blocked or Cloudflare WAF| Wait 5 min (IP block) or set custom User-Agent (SDK)  
405| Method not allowed| Use POST for /v1/messages  
408| Request timeout| Retry — upstream model was slow to respond  
413| Payload too large| Reduce input tokens or message count  
429| Rate limit exceeded (RPM or concurrent)| Check `Retry-After` header; reduce request frequency  
502| Upstream error| Retry after a few seconds  
503| All upstream providers busy| Retry with exponential backoff  
529| Upstream overloaded| Wait and resend — automatic retry recommended  
```

--------------------------------

### Create Message (Gemini)

Source: https://ezaiapi.com/docs

Send a message to the Gemini-2.5-pro model using the EzAI API. Responses are normalized to the Anthropic Messages API format.

```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "gemini-2.5-pro",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Hello, Gemini!" }
    ]
  }'
```

--------------------------------

### Create Message (Claude)

Source: https://ezaiapi.com/docs

Send a message to the Claude-sonnet-4-5 model using the EzAI API. The proxy handles format conversion.

```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Hello, Claude!" }
    ]
  }'
```

--------------------------------

### Uninstall EzAI Proxy Configuration

Source: https://ezaiapi.com/docs

Commands to remove the environment variables and Claude Code settings related to EzAI.

```bash
# Remove env vars from your shell config
unset ANTHROPIC_BASE_URL
unset ANTHROPIC_API_KEY

# Remove Claude Code settings
rm ~/.claude/settings.json
```

--------------------------------

### POST /v1/messages

Source: https://ezaiapi.com/docs

Send a message to any supported AI model. The API automatically handles format conversions for Anthropic, OpenAI, and Google models.

```APIDOC
## Create Message
POST `/v1/messages`

Send a message to any model. The proxy auto-converts between formats upstream — you always use the same Anthropic Messages format.

### Request Example (Claude)
```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{ 
    "model": "claude-sonnet-4-5",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Hello, Claude!" }
    ]
  }'
```

### Request Example (GPT)
```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{ 
    "model": "gpt-4.1",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Hello, GPT!" }
    ]
  }'
```

### Request Example (Gemini)
```bash
curl https://ezaiapi.com/v1/messages \
  -H "x-api-key: sk-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{ 
    "model": "gemini-2.5-pro",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Hello, Gemini!" }
    ]
  }'
```

**Note:** Request uses Anthropic format. The proxy auto-converts to OpenAI format upstream and normalizes the response back.

Response follows the standard Anthropic Messages API format for all models.
```

--------------------------------

### View Rate Limit Response Headers

Source: https://ezaiapi.com/docs

API responses include these headers to track RPM and concurrent request usage.

```text
x-ratelimit-tier: topup
x-ratelimit-limit-requests: 30          # RPM limit for your tier
x-ratelimit-remaining-requests: 28      # Remaining requests this minute
x-ratelimit-reset-requests: 2026-03-03T09:00:00.000Z
x-concurrent-limit: 5                   # Max parallel requests
x-concurrent-remaining: 3               # Available parallel slots
```

--------------------------------

### Authentication Headers

Source: https://ezaiapi.com/docs

Required headers for authenticating requests to the EzAI API.

```text
# Required headers
x-api-key: sk-your-api-key
content-type: application/json
anthropic-version: 2023-06-01
```

--------------------------------

### Count Tokens

Source: https://ezaiapi.com/docs

Endpoint to count the number of tokens in a given message. This operation is free and does not consume credits.

```APIDOC
## POST /v1/messages/count_tokens

### Description
Counts the tokens in a given message. This is a free operation.

### Method
POST

### Endpoint
`/v1/messages/count_tokens`

### Request Body
- **model** (string) - Required - The model to use for counting tokens.
- **messages** (array) - Required - An array of message objects, each with a `role` and `content`.
  - **role** (string) - Required - The role of the message sender (e.g., "user", "assistant").
  - **content** (string) - Required - The content of the message.

### Request Example
```json
{
  "model": "claude-sonnet-4-5",
  "messages": [
    { "role": "user", "content": "Hello!" }
  ]
}
```

### Response
#### Success Response (200)
- **input_tokens** (integer) - The number of input tokens counted.

#### Response Example
```json
{
  "input_tokens": 12
}
```
```

--------------------------------

### OpenAI Responses API

Source: https://ezaiapi.com/docs

An API endpoint that follows the OpenAI Responses format, compatible with tools like n8n and LangChain.

```APIDOC
## POST /v1/responses

### Description
OpenAI Responses API format. Compatible with n8n, LangChain, and tools using the newer OpenAI SDK. Accepts `input` (string or message array) + `instructions` (system prompt). Auto-converted to Anthropic format upstream.

### Method
POST

### Endpoint
`/v1/responses`

### Parameters
#### Request Body
- **model** (string) - Required - The model to use.
- **input** (string or array) - Required - The input prompt or message array.
- **max_output_tokens** (integer) - Optional - The maximum number of tokens to generate.
- **instructions** (string) - Optional - The system prompt.

### Request Example
```json
{
  "model": "claude-sonnet-4-5",
  "input": "Hello!",
  "max_output_tokens": 1024
}
```
```

--------------------------------

### Extended Thinking (Chain-of-Thought)

Source: https://ezaiapi.com/docs

Enable Claude's chain-of-thought reasoning for complex tasks by setting the `thinking` parameter in the request body.

```APIDOC
## Extended Thinking
Enable Claude's chain-of-thought reasoning for complex tasks:

```json
{
  "model": "claude-opus-4-6",
  "max_tokens": 16000,
  "stream": true,
  "thinking": {
    "type": "enabled",
    "budget_tokens": 10000
  },
  "messages": [
    { "role": "user", "content": "Solve this step by step..." }
  ]
}
```

#### Supported Models
claude-opus-4-6 · claude-sonnet-4-5 · claude-sonnet-4

#### Best Practices
budget_tokens: 5K–10K for most tasks. Max: 32,000. Always use with streaming.
```

--------------------------------

### Authentication

Source: https://ezaiapi.com/docs

All requests to the EzAI API require authentication using your API key, which should be included in the request headers.

```APIDOC
## Authentication
Include your API key in every request:

```http
# Required headers
x-api-key: sk-your-api-key
content-type: application/json
anthropic-version: 2023-06-01
```

**Base URL:** `https://ezaiapi.com` — all requests go here. Get your key from the dashboard.
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.