React

Integrating with Alan AI

To integrate a React app with Alan AI:

1. Add the Alan AI Web SDK to your app. You can do it in two ways:

   • With the npm package

   • In the browser library mode

   NPM package

   To install the Alan AI Web SDK with the npm package:

   1. In the Terminal, run the command:

      Terminal

      npm install @alan-ai/alan-sdk-web --save
      

   2. Import alanBtn to your React component:

      Client app

      import alanBtn from "@alan-ai/alan-sdk-web";
      

   Browser library mode

   To load the Alan AI Web SDK in the browser library mode, add the alan_lib.min.js library to the index.html file using the <script> tag:

   Client app

   <script type="text/javascript" src="https://studio.alan.app/web/lib/alan_lib.min.js"></script>
   

2. Add the Alan AI button to your React component:

   • React functional component

   • React class component

   Functional component

   Client app

   import React, { useEffect } from 'react';
   
   useEffect(() => {
     alanBtn({
       key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
       host: 'v1.alan.app',
       onCommand: (commandData) => {
         if (commandData.command === 'go:back') {
           // Call the client code that will react to the received command
         }
       }
     });
   }, []);
   

   If you want to access the component state (or props) inside the callbacks that are passed to the AlanButton, please remember that you may face the «stale closure» problem that occurs because of the closure. To avoid this problem, you may use refs to store your state as shown in the example below:

   Client app

   function useStateReference(value) {
     const ref = useRef(value);
     const [, forceRender] = useState(false);
   
     function updateState(newState) {
       if (!Object.is(ref.current, newState)) {
         ref.current = newState;
         forceRender(s => !s);
       }
     }
   
     return [ref, updateState];
   }
   
   function App() {
     const [count, setCount] = useStateReference({});
     useEffect(() => {
       alanBtn({
         key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
         host: 'v1.alan.app',
         onCommand: () => {
           console.info('Actual count value:', count.current);
         }
       });
     }, []);
   
     return (
       <div className="App">
       </div>
     );
   }
   

   With the code above, the count variable will hold the «actual» value.

   Class component

   Client app

   componentDidMount() {
     this.alanBtnInstance = alanBtn({
       key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
       host: 'v1.alan.app',
       onCommand: (commandData) => {
         if (commandData.command === 'go:back') {
           // Call the client code that will react to the received command
         }
       },
     });
   }
   

Note

Regularly update the @alan-ai/alan-sdk-web package your project depends on. To check if a newer version is available, run npm outdated. To update the alan package, run npm update @alan-ai/alan-sdk-web. For more details, see npm documentation.

Specifying the Alan AI button parameters

You can specify the following parameters for the Alan AI button added to your app:

Name	Type	Description
key	string	The Alan AI SDK key for a project in Alan AI Studio.
host	string	The Alan AI Studio host to which the AI assistant button must connect, for example: 'v1.alan.app'.
authData	JSON object	The authentication or configuration data to be sent to the dialog script. For details, see authData.
rootEl	HTMLElement	The element where Alan AI button will be added. If no rootEl is provided, the Alan AI button is added to the body.
showOverlayOnMicPermissionPrompt	Boolean	The property signaling whether the overlay fade effect must be used when the microphone permission prompt in the browser is displayed. The overlay effect makes the prompt more noticeable and helps make sure users provide microphone access to the AI agent.
onCommand	function	A callback for handling commands from the dialog script. In this callback, you can set up logic on how your app must react to commands received from the dialog script. For details, see onCommand handler.
onButtonState	function	A callback for receiving the connection state of the Alan AI button. For details, see onButtonState handler.
onConnectionStatus	function	A callback for receiving the connection state of the AI agent project run in the Alan AI Cloud. For details, see onConnectionStatus handler.
onEvent	function	A callback responsible for handling events received from Alan AI. For details, see onEvent handler.

Changing the Alan AI button position

By default, the Alan AI button is placed in the bottom right corner of the window. To change the Alan AI button position, you can use the following options for the alanBtn variable:

• left: sets the Alan AI button position from the left edge

• right: sets the Alan AI button position from the right edge

• top: sets the Alan AI button position from the top edge

• bottom: sets the Alan AI button position from the bottom edge

• zIndex: sets the z-order of the Alan AI button

Client app

alanBtn({
   key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
   bottom: '50px',
   left: '50px',
   zIndex: 10
});

Using client API methods

You can use the following client API methods in your React app:

• setVisualState()

• callProjectApi()

• playText()

• sendText()

• playCommand()

• activate()

• deactivate()

• isActive()

• remove()

• textChat.isAudioOutputEnabled()

• textChat.setAudioOutputEnabled()

• textChat.setFullScreenMode()

setVisualState()

Use the setVisualState() method to inform the AI agent about the app’s visual context. For details, see setVisualState().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.setVisualState({ data: 'your data' });
      }}>Send visual state</button>
    </div>
  );
}

export default App;

callProjectApi()

Use the callProjectApi() method to send data from the client app to the dialog script and trigger activities without voice and text commands. For details, see callProjectApi().

Dialog script

projectAPI.setClientData = function(p, param, callback) {
  console.log(param);
};

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.callProjectApi("setClientData", {value:"your data"}, function (error, result){
          // handle error and result here
        });
      }}>Call setClientData method</button>
    </div>
  );
}

export default App;

playText()

Use the playText() method to play specific text in the client app. For details, see playText().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.playText("Hi there, I am Alan");
      }}>Play text</button>
    </div>
  );
}

export default App;

sendText()

Use the sendText() method to send a text message to Alan AI as the user’s input. For details, see sendText().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.sendText("Hello Alan, can you help me?");
      }}>Play text</button>
    </div>
  );
}

export default App;

playCommand()

Use the playCommand() method to execute a specific command in the client app. For details, see playCommand().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
      onCommand: (commandData) => {
        if (commandData.command === 'goBack') {
          // Call client code that will react to the received command
        }
      },
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.playCommand({command: "goBack"});
      }}>Play command</button>
    </div>
  );
}

export default App;

activate()

Use the activate() method to activate the Alan AI button programmatically. For details, see activate().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.activate();
      }}>Activate the Alan AI button</button>
    </div>
  );
}

export default App;

deactivate()

Use the deactivate() method to deactivate the Alan AI button programmatically. For details, see deactivate().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.deactivate();
      }}>Deactivate the Alan AI button</button>
    </div>
  );
}

export default App;

isActive()

Use the isActive() method to check the Alan AI button state: active or not. For details, see isActive().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        console.log('Button is active: ' + alanBtnRef.btnInstance.isActive());
      }}>Check the button state</button>
    </div>
  );
}

export default App;

remove()

Use the remove() method to remove the Alan AI button from the parent element. For details, see remove().

Client app

import alanBtn from '@alan-ai/alan-sdk-web';
import React, { useEffect, useRef } from 'react';

function App() {
  const alanBtnRef = useRef({}).current;
  useEffect(() => {
    alanBtnRef.btnInstance = alanBtn({
      key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
      host: 'v1.alan.app',
    });
  }, []);

  return (
    <div className="App">
      <button onClick={() => {
        alanBtnRef.btnInstance.remove();
      }}>Remove the Alan AI button</button>
    </div>
  );
}

export default App;

textChat.isAudioOutputEnabled()

Use the textChat.isAudioOutputEnabled() method to get the state of audio output for the Alan AI Chat. For details, see textChat.isAudioOutputEnabled().

Client app

alanBtnInstance.textChat.isAudioOutputEnabled()

textChat.setAudioOutputEnabled()

Use the textChat.setAudioOutputEnabled() method to enable or disable audio output for the Alan AI Chat. For details, see textChat.setAudioOutputEnabled().

Client app

alanBtnInstance.textChat.setAudioOutputEnabled(true)

textChat.setFullScreenMode()

Use the textChat.setFullScreenMode() method allows switching the Alan AI Chat interface between full-screen and normal view. For details, see textChat.setFullScreenMode().

Client app

alanBtnInstance.textChat.setFullScreenMode(true);

Using handlers

You can use the following Alan AI handlers in your React app:

• onCommand handler

• onButtonState handler

• onConnectionStatus handler

• onEvent handler

onCommand handler

Use the onCommand handler to handle commands sent from the dialog script. For details, see onCommand handler.

Client app

alanBtn({
  key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
  host: 'v1.alan.app',
  onCommand: (commandData) => {
    if (commandData.command === 'go:back') {
      // Call client code that will react to the received command
    }
  },
});

onButtonState handler

Use the onButtonState handler to capture and handle the Alan AI button state changes. For details, see onButtonState handler.

Client app

alanBtn({
  key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
  host: 'v1.alan.app',
  onButtonState: function (e) {
    console.info('onButtonState', e);
  },
});

onConnectionStatus handler

Use the onConnectionStatus handler to capture and handle the connection status for the AI agent project. For details, see onConnectionStatus handler.

Client app

alanBtn({
  key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
  host: 'v1.alan.app',
  onConnectionStatus: function(status) {
    console.log("The status is " + status);
  },
});

onEvent handler

Use the onEvent handler to capture and handle events emitted by Alan AI: get user’s utterances, agent responses and so on. For details, see onEvent handler.

Client app

alanBtn({
  key: 'YOUR_KEY_FROM_ALAN_STUDIO_HERE',
  host: 'v1.alan.app',
  onEvent: function (e) {
    console.info('onEvent', e);
  },
});

What’s next?

Example apps

Find and explore examples of AI agents in Alan AI GitHub.