A lightweight chatbot library. (NOTE: This library needs the user to be connected online to provide results.)
Check out the example at: AetherLexLib AI Chatbot
This project is under "MIT LICENSE" so it is free and you can use it, modify it etc. but only if you included the copy of the license of the project you used.
To get the library do one of these methods:
-
Downloading latest update here
-
Run
npm i aetherlexlib_jsto dowload it using npm. -
Link it to your html using:
<script src="https://cdn.skypack.dev/aetherlexlib_js"></script>
After you've downloaded the .zip, extract it and add the "project_build" and "LICENSE" files to your project to start using it.
Step 1:
- Add a script tag with type "module" in your HTML.
Example:<script type="module"> // Your JavaScript code will go here </script>
Step 2:
- Import the library at the beginning of your script.
Example:import { analyzeAndRespond } from './path/to/your/library.js';
Step 3:
- Prepare your conversation data. Ensure you have a structured dataset for training your AI. You can use a JSON format or text files as necessary.
Step 4:
- Call the
analyzeAndRespondfunction with user input.
Example:const userInput = "Hello, how are you?"; const response = await analyzeAndRespond(userInput); console.log(response); // Outputs the AI's response
Step 5:
- Handle responses appropriately within your application. You can display the response in your user interface or process it further as needed.
Step 6:
- (Optional) Customize the predefined responses and learning data to fit your specific use case. Modify the arrays in the
responsesobject to add more context and variation.
Here’s a simple example that ties everything together:
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Chatbot Example</title>
</head>
<body>
<div id="chat-container"></div>
<input type="text" id="user-input" placeholder="Type your message..." />
<button id="send-button">Send</button>
<script type="module">
import { analyzeAndRespond } from './path/to/your/library.js';
document.getElementById('send-button').addEventListener('click', async () => {
const inputField = document.getElementById('user-input');
const userInput = inputField.value;
const response = await analyzeAndRespond(userInput);
// Display the response
const chatContainer = document.getElementById('chat-container');
chatContainer.innerHTML += `<div>User: ${userInput}</div>`;
chatContainer.innerHTML += `<div>AI: ${response}</div>`;
inputField.value = ''; // Clear the input field
});
</script>
</body>
</html>An example on how to use the engine can be found on "examples" folder, you can test it here: AetherLexLib AI Chatbot
If you are unable to use the library via cdn or npm, you can use the AetherLexAPI directly. This is a simple API that you can use to interact with the AI.
const message = "hello";
const response = await fetch('https://aetherlexlib-js.onrender.com/api/analyze', {
method: 'POST',
headers: {
'Content-Type': 'application/json',
},
body: JSON.stringify({ analyzedResult:message }),
});
if (!response.ok) {
throw new Error(`Error: ${response.status}`);
}
const data = await response.json();
responseDiv.innerHTML = `<p>Response: ${JSON.stringify(data.processedResult)}</p>`;
Thats it! with this, you can send and recieve messages from the AI.
The body is the most important thing of sending requests and it is also the one that can cause you problems!
You need to exactly add analyzedResult: as the message key and the message as the value. If you do not do this, the AI will not respond and throw and error telling to use analyzedResult:
The response the AI will give, will be in JSON form, but for some reason if you need to just get the response as plain text and not JSON, just add this in the body element along with the analyzedResult:
body: JSON.stringify({ analyzedResult:message, isJson: false })
If isJson is false, then the API would only respond with plain text.
Example:
const response = await fetch('https://aetherlexlib-js.onrender.com/api/analyze', {
method: 'POST',
headers: {
'Content-Type': 'application/json',
},
body: JSON.stringify({ analyzedResult:message, isJson: false }),
});
if (!response.ok) {
throw new Error(`Error: ${response.status}`);
}
const data = await response.text();
responseDiv.innerHTML = `<p>Response: ${data}</p>`;
- This chatbot library uses the "Fuse" library which is licensed under "Apache 2.0".
- For more knowledge, the AI also uses WikiData which is licensed under CC0.
- Additionally, the AI utilizes DuckDuckGo for supplementary search data (DuckDuckGo under its own terms).
- This library incorporates the "Compromise" library (Licensed under MIT).