HUGJS - Huggingface Api client

A lightweight JavaScript client for interacting with the Hugging Face chat/completions API.

npm install hug-client

#USAGE 1: index.js
#Self calling function for on page start or refresh executions

import Hug from "./hug.js";

let apiToken = "YOUR HUGGINGFACE ACCESS TOKEN";

(async ()=>{
    const resp = await Hug({ 
    messages: [
        {role: "user", content: "Where was Okomfo Anokye born" //userInput,},
        {role: "system", // THIS OBJECT is  OPTIONALcontent: "Chat Bot" //instruction on What The model should do}
    ],
    model: "meta-llama/Llama-3.1-8B-Instruct:cerebras"//Model Name,
}, apiToken);

console.log(resp);
})();


## USAGE 2: index.js
# Event 

const apiToken = <Your Access token>

const generate = async ()=>{
    const resp = await Hug({
        messages: [
            {role: "user", content: userInput.value},
            {role: "system", content: "ChatBot AI"},
        ],
        model: "meta-llama/Llama-3.1-8B-Instruct:cerebras"
    }, api);

    res.textContent = resp;
    
};

btn.addEventListener('click',()=>{
    generate();
    console.log("clicked")
})



API Reference
Hug(data, apiToken);
- data: Object containing the request payload

- messages: Array of { role, content } objects

- role: "user", "system", or "assistant"

- content: String message content

- model: String name of the Hugging Face model to query

- apiToken: Your Hugging Face API key (hf_xxx...) ## Visit: Huggingface.co site to find api/access token

Returns: A Promise resolving to the JSON response from Hugging Face.

Example payload:

{
  "messages": [
    { "role": "user", "content": "Hello!" }
  ],
  "model": "meta-llama/Llama-3-8b-chat-hf" # Visit: Huggingface.co site to find models
}


Contributors: Prince Adu. Boahene