[go: nahoru, domu]

Skip to content

Latest commit

ย 

History

History
209 lines (152 loc) ยท 11.5 KB

README.md

File metadata and controls

209 lines (152 loc) ยท 11.5 KB

Character AI Unofficial Node API

Node.js client for the unofficial Character AI API, an awesome website which brings characters to life with AI!

Intro

This repository is inspired by RichardDorian's unofficial node API. Though, I found it hard to use and it was not really stable and archived. So I remade it in javascript.

This project is not affiliated with Character AI in any way! It is a community project. The purpose of this project is to bring and build projects powered by Character AI.

If you like this project, please check their website.

Features

  • ๐Ÿ‘ Fully written in Javascript and CommonJS (for max compatibility and ease of use)
  • โŒš Asynchronous requests
  • ๐Ÿ—ฃ๏ธ Use conversations or use the API to fetch information
  • ๐Ÿงธ Easy to use
  • ๐Ÿ” Active development
  • ๐Ÿ‘ค Guest & token login support

Installation

npm install node_characterai

Usage

Basic guest authentication and message:

const CharacterAI = require("node_characterai");
const characterAI = new CharacterAI();

(async () => {
  // Authenticating as a guest (use `.authenticateWithToken()` to use an account)
  await characterAI.authenticateAsGuest();

  // Place your character's id here
  const characterId = "8_1NyR8w1dOXmI1uWaieQcd147hecbdIK7CeEAIrdJw";

  // Create a chat object to interact with the conversation
  const chat = await characterAI.createOrContinueChat(characterId);

  // Send a message
  const response = await chat.sendAndAwaitResponse("Hello discord mod!", true);

  console.log(response);
  // Use `response.text` to use it as a string
})();

Using an Access Token

Some parts of the API (like managing a conversation) require you to be logged in using a sessionToken.

To get it, you can open your browser, go to the Character.AI website in localStorage.

Important

If you are using old versions of the package and are getting a Authentication token is invalid, you now again need a sessionToken to authenticate (as of update 1.2.5 and higher). See below.

If you are using something that is using the package and has not updated to the latest version in a while, make sure to update the package by doing npm update node_characterai or manually copying the files or open a respective issue to their package (if they have one).


โš ๏ธ WARNING: DO NOT share your session token to anyone you do not trust or if you do not know what you're doing.

Anyone with your session token could have access to your account without your consent. Do this at your own risk.


On PC:

  1. Open the Character.AI website in your browser (https://beta.character.ai)
  2. Open the developer tools (F12, Ctrl+Shift+I, or Cmd+J)
  3. Go to the Application tab
  4. Go to the Storage section and click on Local Storage
  5. Look for the char_token key
  6. Open the object, right click on value and copy your session token.

Session_Token

On Mobile:

  1. Open the Character.AI website in your browser (https://beta.character.ai)
  2. Open the URL bar, write javascript: (case sensitive) and paste the following:
(function(){let e=window.localStorage["char_token"];if(!e){alert("You need to log in first!");return;}let t=JSON.parse(e).value;document.documentElement.innerHTML=`<div><i><p>provided by node_characterai - <a href="https://github.com/realcoloride/node_characterai?tab=readme-ov-file#using-an-access-token">click here for more information</a></p></i><p>Here is your session token:</p><input value="${t}" readonly><p><strong>Do not share this with anyone unless you know what you are doing! This is your personal session token. If stolen or requested by someone you don't trust, they could access your account without your consent; if so, please close the page immediately.</strong></p><button id="copy" pl-s1">${t}'); alert('Copied to clipboard!')">Copy session token to clipboard</button><button >Refresh the page</button></div>`;localStorageKey=null;storageInformation=null;t=null;})();
  1. The following page should appear: Access_Token_Mobile

  2. Click the respective buttons to copy your access token or id token to your clipboard.


When using the package, you can:

  • Login as guest using authenticateAsGuest() - for mass usage or testing purposes
  • Login with your account or a token using authenticateWithToken() - for full features and unlimited messaging

Finding your character's ID

You can find your character ID in the URL of a Character's chat page.

For example, if you go to the chat page of the character Discord Moderator you will see the URL https://beta.character.ai/chat?char=8_1NyR8w1dOXmI1uWaieQcd147hecbdIK7CeEAIrdJw.

The last part of the URL is the character ID: Character_ID

Image Interactions

WARNING: This part is currently experimental, if you encounter any problem, open an Issue.

๐Ÿ–ผ๏ธ Character AI has the ability to generate and interpret images in a conversation. Some characters base this concept into special characters, or maybe use it for recognizing images, or to interact with a character and give it more details on something: the possibilities are endless.

๐Ÿ’ Most of the Character AI image features can be used like so:

// Most of these functions will return you an URL to the image
await chat.generateImage("dolphins swimming in green water");
await chat.uploadImage("https://www.example.com/image.jpg");
await chat.uploadImage("./photos/image.jpg");

// Other supported types are Buffers, Readable Streams, File Paths, and URLs
await chat.uploadImage(imageBuffer);

// Including the image relative path is necessary to upload an image
await chat.sendAndAwaitResponse({
  text: "What is in this image?",
  image_rel_path: "https://www.example.com/coffee.jpg",
  image_description: "This is coffee.",
  image_description_type: "HUMAN" // Set this if you are manually saying what the AI is looking at
}, true);

Props to @creepycats for implementing most of this stuff out

Troubleshooting

Problem Answer
โŒ Token was invalid Make sure your token is actually valid and you copied your entire token (its pretty long) or, you have not updated the package.
โš ๏ธ The specified Chromium path for puppeteer could not be located On most systems, puppeteer will automatically locate Chromium. But on certain distributions, the path has to be specified manually. This warning occurs if node_characterai could not locate Chromium on linux (/usr/bin/chromium-browser), and will error if puppeteer cannot locate it automatically. See this for a fix.
๐Ÿ˜ฎ Why are chromium processes opening? This is because as of currently, the simple fetching is broken and I use puppeteer (a chromium browser control library) to go around cloudflare's restrictions.
๐Ÿ‘ฅ authenticateAsGuest() doesn't work See issue #14.
๐Ÿฆ’ Hit the max amount of messages? Sadly, guest accounts only have a limited amount of messages before they get limited and forced to login. See below for more info ๐Ÿ‘‡
๐Ÿช How to use an account to mass use the library? You can use conversations, a feature introduced in 1.0.0, to assign to users and channels. To reproduce a conversation, use OOC (out of character) to make the AI think you're with multiple people. See an example here: chrome_RDbmXXtFNlchrome_BgF8crPvqC (Disclaimer: on some characters, their personality will make them ignore any OOC request).
๐Ÿ“ฃ Is this official? No, this project is made by a fan of the website and is unofficial. To support the developers, please check out their website.
๐Ÿ˜ฒ Did something awesome with node_characterai? Please let me know!
โœ‰๏ธ Want to contact me? See my profile
๐Ÿ’ก Have an idea? Open an issue in the Issues tab
โž• Other issue? Open an issue in the Issues tab

In-depth troubleshooting

๐Ÿคš Before you scroll, please know that:

  • In the Client class, you can access the Requester class and define puppeteer or other variables related to how CharacterAI will work using characterAI.requester.(property). Also, anything here is subject to change, so make sure to update the package frequently.

๐Ÿ’ณ About CharacterAI+

"I am a member of cai+, how do I use it?"

Change the property .usePlus from the requester and if needed, change .forceWaitingRoom.

For example:

// Default is `false`
characterAI.requester.usePlus = true;

๐Ÿงญ About Puppeteer

Around a few months ago, the package only required the node-fetch module to run. The package was made using simple API requests.

However, over time, Cloudflare started fighting against scraping and bots, which also made it almost impossible for anyone to use this package.

This is where in versions 1.1 and higher, puppeteer is used (which uses a chromium browser) to make requests with the API.

โš™๏ธ How to change Puppeteer settings

๐Ÿ‘‰ IMPORTANT: do the changes before you initialize your client!

In the CharacterAI class, you can access the requester and define the .puppeteerPath variable or other arguments, and the properties include (and are subject to change in future versions):

// Chromium executable path (in some linux distributions, /usr/bin/chromium-browser)
puppeteerPath;
// Default arguments for when the browser launches
puppeteerLaunchArgs;
// Boolean representing the default timeout (default is 30000ms)
puppeteerNoDefaultTimeout;
// Number representing the default protocol timeout
puppeteerProtocolTimeout;
Specifying Chromium's path

๐Ÿง For linux users, if your puppeteer doesn't automatically detect the path to Chromium, you will need to specify it manually.

To do this, you just need to set puppeteerPath to your Chromium path:

characterAI.puppeteerPath = "/path/to/chromium-browser";

On Linux, you can use the which command to find where Chromium is installed:

$ which chromium-browser # or whatever command you use to launch chrome

๐Ÿ’ก I recommend that you frequently update this package for bug fixes and new additions.

Disclaimer

โค๏ธ This project is updated frequently, always check for the latest version for new features or bug fixes.

๐Ÿš€ If you have an issue or idea, let me know in the Issues section.

๐Ÿ“œ If you use this API, you also bound to the terms of usage of their website.

(real)coloride - 2023-2024, Licensed MIT.