gpt4free-client
v2.0.7
Published
A powerful Node.js script to communicate with big text AI models (like gpt4 or OpenChat)
Downloads
26
Maintainers
Readme
Gpt4free-Client
The Gpt4free-Client is a Node.js library that allows you to interact with many artificial intelligence models for free text generation. This library is inspired by gpt4free, developed by @xtekky and created by @Anybody.
Installation
npm install Gpt4free-Client
Usage
const GptClient = require('Gpt4free-Client');
async function generateText() {
// Create chat completion
const completion = await GptClient.create({
model: 'gpt-4-0613',
temperature: 0.7,
systemPrompt: 'You are powerful AI assistant\n',
messages: [
{
role: 'system',
messages: 'hello world'
}
]
});
// Get generated text
const generatedText = await completion;
console.log(generatedText);
}
generateText();
API
GptClient.create(options)
Creates a chat completion with the specified options.
- model: optional (default='gpt-4-0613'), the name of the AI model to use.
- temperature: optional (default=0.7), the temperature for generating responses.
- systemPrompt: optional (default='You are powerful AI assistant'), the system prompt to use.
- messages: a list of messages for the chat. Each message is an object with role (either 'system' or 'user') and message properties.
- stream: optional (default=false), whether to use streaming mode for the response.
Returns a Promise that resolves to the generated answer from the model.
API usage
Using a different model
const completion = await GptClient.create({
model: 'openchat',
temperature: 0.7,
systemPrompt: 'You are powerful AI assistant\n',
messages: [
{
role: 'system',
messages: 'hello world'
}
]
});
console.log(completion); // answer
Setting new temperature
const completion = await GptClient.create({
model: 'gpt-4-0613',
temperature: 1,
systemPrompt: 'You are powerful AI assistant\n',
messages: [
{
role: 'system',
messages: 'hello world'
}
]
});
console.log(completion); // answer
Using custom system prompt
const completion = await GptClient.create({
model: 'gpt-4-0613',
temperature: 0.7,
systemPrompt: 'You're AI copy of DaniilSV, the best brawl stars mods creator',
messages: [
{
role: 'system',
messages: 'hello world'
}
]
});
console.log(completion); // answer
Using context mode
const completion = await GptClient.create({ model: 'gpt-4-0613', temperature: 0.7, systemPrompt: 'You are powerful AI assistant\n', messages: [ { role: 'system', messages: 'what is Gpt-Clinet?' }, { role: 'system', messages: 'repeat your answer' }, { role: 'system', messages: 'Thanks you!' } ] });
console.log(completion); // answer
#### Using streaming mode
```JavaScript
const completionStream = await GptClient.create({
model: 'gpt-4-0613',
temperature: 0.7,
systemPrompt: 'You are powerful AI assistant\n',
messages: [
{
role: 'system',
messages: 'hello world'
}
],
stream: true
});
completionStream.on('data', (chunk) => {
console.log(chunk);
});
completionStream.on('end', () => {
console.log('Completed!');
});
You can also find more API usages in tests
folder
Contributing
Contributions are welcome! If you have any ideas, suggestions, or bug reports, please open an issue or submit a pull request.