Skip to content
Switch branches/tags
Go to file

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Use's natural language tools in your Botkit-powered Bot!

This middleware plugin for Botkit allows you to seamlessly integrate natural language intent APIs into your Botkit bot. provides a service that uses machine learning to help developers handle natural language input. The Wit API receives input from the user, and translates it into one or more "intents" which map to known actions or choices. The power of Wit is that it can continually be trained to understand more and more responses without changing the underlying bot code!


Create a Wit application here. Then, set up and train at least one intent.

From your app's settings page, snag the Server Access Token. You will need this to use Wit's API.

Add botkit-middleware-witai as a dependency to your Botkit bot!

npm install --save botkit-middleware-witai

Enable the middleware:

var wit = require('botkit-middleware-witai')({
    token: <my_wit_token>


controller.hears(['hello'],'direct_message',wit.hears,function(bot, message) {

    // ...

For a full example example_bot.js

What it does

Using the Wit receive middleware with Botkit causes every message that is sent to your bot to be first sent to for processing. The results of the call to are added into the incoming message as message.entities, and will match the results of this API call.

Using the Wit hears middleware tells Botkit to look for intents information, and match using this information instead of the built in pattern matching function.

You must make an intent entity in the understandings area of and train it to register certain expressions.

I.e "intent" -> "weather"

Expression: "What is the weather?" and that maps to the weather intent.

Unless you want to directly access the information returned by wit, you can use this transparently by enabling bot the receive and hears middlewares.


Middleware for using with Botkit-powered bots




No releases published


No packages published