Cheshire Cat AI API Client

API Client made in TypeScript to communicate with the Cheshire Cat AI.
The package provides a class to interface with the Cheshire Cat AI backend.
It can be used both in Browser and NodeJS environments.

Every endpoint is a CancelablePromise, which means you can cancel the request if you want.

Installation

npm install ccat-api
# OR
yarn add ccat-api
# OR
pnpm i ccat-api

Getting started

To set up the client, you must first import the CatClient class:

import { CatClient } from 'ccat-api'

const cat = new CatClient({
    baseUrl: 'localhost',
    userId: 'user'.
    //... other settings
})

cat.send('Hello from a user!') // this will send a message to the /ws/user

cat.userId = 'new_user'

cat.send('Hello from a new user!') // this will send a message to the /ws/new_user

Client settings

API_KEY, CORE_HOST, CORE_PORT and CORE_USE_SECURE_PROTOCOLS refer to the CCAT Core .env file.

Property	Type	Default	Description
baseUrl	string	Required	The same of CORE_HOST
authKey	string	''	The same of API_KEY
port	number	1865	The same of the CORE_PORT
secure	boolean	false	The same of the CORE_USE_SECURE_PROTOCOLS
user	string	'user'	The user ID to use for the WebSocket and the API client
instant	boolean	true	Instantly initialize the WebSocket and the API client, or later with .init()
timeout	number	10000	Timeout for the endpoints, in milliseconds
headers	object	{}	The headers to send with the API requests
ws	string	undefined	An object of type WebSocketSettings

WebSocket settings

Property	Type	Default	Description
path	string	'ws'	Websocket path to use to communicate with the CCat
retries	number	3	The maximum number of retries before calling onFailed
query	string	undefined	The query to send with the WebSocket connection
delay	number	3000	The delay for reconnect, in milliseconds
onFailed	function	undefined	The function to call after failing all the retries

Then, for example, you can configure the LLM like this:

cat.api.settingsLargeLanguageModel.upsertLlmSetting('LLMOpenAIConfig', {
    openai_api_key: 'OPEN_API_KEY'
})

To send a message to the cat, you can:

cat.send('Hello my friend!')

You can listen to the WebSocket events:

cat.onConnected(() => {
    console.log('Socket connected')
}).onMessage(msg => {
    console.log(msg)
}).onError(err => {
    console.log(err)
}).onDisconnected(() => {
    console.log('Socket disconnected')
})

For example, you can get the list of plugins like this:

cat.api.plugins.listAvailablePlugins().then(plugins => {
    console.log(plugins)
})

Credits

Made for the Cheshire Cat AI organization.

This was possible thanks to openapi-typescript-codegen.