Chat bots can make tools on the web more accessible, especially when embedded on websites or mobile apps. One of these tools is Wolfram Alpha, which is one of the web's strongest computational knowledge engines. Wolfram Alpha is capable of answering many different types of queries, from retrieving a picture of a cat for you to solving systems of equations or difficult calculus questions. Here's an example, feel free to chat and interact with our Wolfram Alpha chat bot here.
Creating your own Wolfram Alpha chat bot is as simple as a couple clicks, by using our Wolfram Alpha chat bot template, which has already been connected to Wolfram Alpha for you! First, create an account on botlibre.com or botlibre.biz. Click on Bots (in the header under the BOTlibre! logo), and select the option to create a New Bot:
Then in the template section, type: wolfram-alpha-bot-template as shown below:
Enter your other parameters and Create the bot. Now you have your very own copy of the Wolfram Alpha chat bot! You can edit your bot's behaviour from the My Bots --> <Your bot Name> --> Admin --> Training & Chat Log (for non-programmers) or Scripts (which is in the Self scripting language, which you can learn more about here).
If you want to embed your bot to a website or blog, simply use the My Bots --> <Your bot Name> --> Embed.
If you're interested in the bot's functionality or confused on how the bot works:
- In general, any text you message the bot with (except messages like "Hello" and "How are you doing" which will be met with their default responses) will be queried and inputted into the Wolfram Alpha engine, parsing through its responses and returning only the top 1-2 outputs from it (different from the real Wolfram Alpha search engine, which can give you many possible responses). These may take a while to process as they are images, so please be patient.
- No matter what text input you message the bot with, the bot will enter that message character by character as a query to the Wolfram Alpha engine. However, the bot will say "Sorry, I do not understand." if Wolfram Alpha does not understand your input or gives no valid responses.
- Sometimes the bot returns only one image and sometimes it returns two images, why is that? When the bot returns two images, it detects the first output as an "input" or "input interpretation" pod, which in most cases does not include the results of the query but may include important clarifications on how your input was interpreted by the Wolfram Alpha engine. In this case, it is necessary to also return the second output from Wolfram Alpha, as it contains the result of the search that the user was most likely looking for, with the first output, for any clarification. When the bot returns one image, the input was included in this output in some way (such as a math equation), so it serves as both the input interpretation and the result.
- If the bot outputs a result you weren't looking for, chances are that the result you were looking for was not the top result. It may then be helpful to add some keywords to guide the Wolfram Alpha engine. For example, if you want to solve the equation x+9=10, if you just input "x+9=10", the bot will return a graph of x+9=10, as x=1 was not the top result. However, if you input "solve x+9=10", the bot will return "x=1", as desired.