Demo: A Simple Voice AI Bot with Web Speech API and Node.js

This is how this web app works:

1. Using the Web Speech API’s SpeechRecognition interface to listen your voice from a microphone
2. Send your message to API.ai (the natural language processing platform) as a text string
3. Once the AI from the API.ai returns the reply text back, use the SpeechSynthesis interface to give it a synthetic voice.

Try It on Your Own Server

Rename the .env_test to .env and fill the env vars:

APIAI_TOKEN=
APIAI_SESSION_ID=

The first one is an API.ai API key (Please get one by sign up with API.ai), and the second one is a session ID, which is an arbitrary string.

Or use this Heroku button to deploy to Heroku server. You just need to fill out the env vars with the API key and a session ID. No need to create an .env file.