An example Alexa Skill to query Wolfram Alpha, written in Python.

• An AWS account
• A (free) Wolfram Alpha API AppID

1. Create an AWS Lambda function using alexa-wolfram-alpha.py as the code
   • You can follow the official Amazon instructions to give you a hand
   • Put your Wolfram Alpha AppID into an environment variable called "WOLFRAM_ID" and also put your Alexa Skill Identifier in an environement variable called "SKILL_ID". For help view the documentation
   • You can use test_event.json as your test template
   • Consider extending the timeout beyond the default of 3 seconds (I raised mine to 10, which is likely excessive, but eliminated some sporadic errors e.g. #1)
2. Create a new Alexa Skill using intent_schema.json and sample_utterances.txt
3. Test that it's working from the web interface during the creation of the skill
4. Test that it's working with your Echo

Feel free to fork and hack on this.

The setup.cfg file is just in case you have a homebrew installed python2 and want to include other 3rd party libraries in your code, in which case you have to upload a zip of your code as the lambda function. To do so, from your root directory do something like:

1. pip2 install LIBRARY_NAME -t .
2. zip -r $(date +%s)-alexa-wolfram-alpha.zip .

Also, take a look at Amazon's official sample code for a lambda-based Alexa Skill in Python, named alexa-skills-kit-color-expert-python. (I found it very helpful to reference but couldn't find an easy way to view the source online, so I copied it to a GitHub Gist.)

• https://github.com/astrobleem/askwolfram :: A similar project in PHP that helped me figure out what the intent schema and sample utterances should look like