index.html<\/b>: The frontend interface (chat UI) where users can interact with the ChatGPT-powered chatbot.<\/li>\n<\/ul>\n\ud83d\ude80 Let’s get started with ChatGPT API Integration in PHP step by step, using a ready-made and production-ready script.<\/h4>\nAPI Configuration (config.php)<\/h2>\n
The config.php<\/code> file contains essential configuration settings for the ChatGPT API integration. Here, you will define your OpenAI API key, select the model to use, and set other parameters like temperature and maximum tokens.<\/p>\n<?php
<\/span>\/\/ Your OpenAI API Key (get it from https:\/\/platform.openai.com\/api-keys)
<\/span>define<\/span>(<\/span>'OPENAI_API_KEY'<\/span>, <\/span>'your-api-key-here'<\/span>);
<\/span>\/\/ API Endpoint
<\/span>define<\/span>(<\/span>'OPENAI_API_ENDPOINT'<\/span>, <\/span>'https:\/\/api.openai.com\/v1\/chat\/completions'<\/span>);
<\/span>\/\/ Model to use
<\/span>define<\/span>(<\/span>'OPENAI_MODEL'<\/span>, <\/span>'gpt-3.5-turbo'<\/span>); <\/span>\/\/gpt-5.2, gpt-5-nano
\/\/ Request timeout (in seconds)
<\/span>define<\/span>(<\/span>'REQUEST_TIMEOUT'<\/span>, <\/span>30<\/span>);
<\/span>\/\/ Temperature (controls randomness: 0 = deterministic, 1 = random)
<\/span>define<\/span>(<\/span>'TEMPERATURE'<\/span>, <\/span>0.7<\/span>);
<\/span>\/\/ Maximum tokens in response
<\/span>define<\/span>(<\/span>'MAX_TOKENS'<\/span>, <\/span>2000<\/span>);
<\/span>?><\/span><\/pre>\nMake sure to replace 'your-api-key-here'<\/code> with your actual OpenAI API key obtained from the OpenAI platform.<\/i><\/p>\nChatGPT Client Class (ChatGPTClient.php)<\/h2>\n
The ChatGPTClient.php<\/code> file contains a PHP class that encapsulates the logic for communicating with the ChatGPT API. This class handles sending requests and processing responses.<\/p>\nSome of the key features of this class include:<\/p>\n
\n- Initialization with API key, model, timeout, temperature, and max tokens.<\/li>\n
- Building request payloads for the ChatGPT API.<\/li>\n
- Sending HTTP requests to the ChatGPT API via cURL.<\/li>\n
- Error handling and response processing.<\/li>\n
- Support for conversation history and system roles.<\/li>\n<\/ul>\n
Core Method: sendMessage()<\/b>
\nThis method sends a user prompt along with optional conversation history and returns the AI response.
\nIt automatically:<\/p>\n\n- Prepends system prompts<\/li>\n
- Appends user messages<\/li>\n
- Parses the API response<\/li>\n<\/ul>\n
<?php
<\/span>class <\/span>ChatGPTClient
<\/span>{
private <\/span>$apiKey<\/span>;
private <\/span>$endpoint<\/span>;
private <\/span>$model<\/span>;
private <\/span>$timeout<\/span>;
private <\/span>$temperature<\/span>;
private <\/span>$maxTokens<\/span>;
private <\/span>$lastError<\/span>;
<\/span>\/**
* Constructor
*
* @param string $apiKey OpenAI API Key
* @param string $model Model to use (default: gpt-3.5-turbo)
* @param int $timeout Request timeout in seconds
* @param float $temperature Temperature for response (0-2)
* @param int $maxTokens Maximum tokens in response
*\/
<\/span>public function <\/span>__construct<\/span>(
<\/span>$apiKey<\/span>,
<\/span>$model <\/span>= <\/span>'gpt-3.5-turbo'<\/span>,
<\/span>$timeout <\/span>= <\/span>30<\/span>,
<\/span>$temperature <\/span>= <\/span>0.7<\/span>,
<\/span>$maxTokens <\/span>= <\/span>2000
<\/span>) {
<\/span>$this<\/span>-><\/span>apiKey <\/span>= <\/span>$apiKey<\/span>;
<\/span>$this<\/span>-><\/span>model <\/span>= <\/span>$model<\/span>;
<\/span>$this<\/span>-><\/span>endpoint <\/span>= <\/span>'https:\/\/api.openai.com\/v1\/chat\/completions'<\/span>;
<\/span>$this<\/span>-><\/span>timeout <\/span>= <\/span>$timeout<\/span>;
<\/span>$this<\/span>-><\/span>temperature <\/span>= <\/span>$temperature<\/span>;
<\/span>$this<\/span>-><\/span>maxTokens <\/span>= <\/span>$maxTokens<\/span>;
<\/span>$this<\/span>-><\/span>lastError <\/span>= <\/span>null<\/span>;
}
<\/span>\/**
* Send a message to ChatGPT and get a response
*
* @param string $message User message
* @param array $conversationHistory Previous messages for context (optional)
* @return string|false Response from ChatGPT or false on failure
*\/
<\/span>public function <\/span>sendMessage<\/span>(<\/span>$message<\/span>, <\/span>$conversationHistory <\/span>= [])
{
try {
<\/span>\/\/ Prepare messages array
<\/span>$messages <\/span>= <\/span>$conversationHistory<\/span>;
<\/span>$messages<\/span>[] = [
<\/span>'role' <\/span>=> <\/span>'user'<\/span>,
<\/span>'content' <\/span>=> <\/span>$message
<\/span>];
<\/span>\/\/ Prepare request payload
<\/span>$payload <\/span>= [
<\/span>'model' <\/span>=> <\/span>$this<\/span>-><\/span>model<\/span>,
<\/span>'messages' <\/span>=> <\/span>$messages<\/span>,
<\/span>'temperature' <\/span>=> <\/span>$this<\/span>-><\/span>temperature<\/span>,
<\/span>'max_tokens' <\/span>=> <\/span>$this<\/span>-><\/span>maxTokens
<\/span>];
<\/span>\/\/ Send request
<\/span>$response <\/span>= <\/span>$this<\/span>-><\/span>makeRequest<\/span>(<\/span>$payload<\/span>);
if (<\/span>$response <\/span>=== <\/span>false<\/span>) {
return <\/span>false<\/span>;
}
<\/span>\/\/ Extract response content
<\/span>if (isset(<\/span>$response<\/span>[<\/span>'choices'<\/span>][<\/span>0<\/span>][<\/span>'message'<\/span>][<\/span>'content'<\/span>])) {
return <\/span>trim<\/span>(<\/span>$response<\/span>[<\/span>'choices'<\/span>][<\/span>0<\/span>][<\/span>'message'<\/span>][<\/span>'content'<\/span>]);
}
<\/span>$this<\/span>-><\/span>lastError <\/span>= <\/span>'Invalid response format from API'<\/span>