Inspiration
Some of our friends speak too much. Ever woken up in the morning, checked your phone and realised that you have hundreds of unread messages? We can't be bothered to read them - can you?
The reason we developed a piece of software for the Amazon Echo was that this can be queried on the go. We don't need to stop and look at a screen to be updated - we can just listen to it whilst doing other things. Time saving is very important in this day and age, and we want to cater to this by allowing users to do tasks in parallel.
What it does
Our code looks through a specific chat's history and finds the user's last message. From that point on, we assume that the user has not read any messages. From these messages, we select the lines of text for Alexa to say in the order of "importance" (this can be changed depending on what the user finds important).
How we built it
We wrote a lambda function instead of using web services to host the server. Customised intents, utterances and implementations of Alexa skills (written in nodejs) were then written using the Amazon Alexa api.
Challenges we ran into
Getting Alexa to say "hello world" was a challenging task (it took us 8 hours). Part of this required knowledge of writing a lambda function and linking this to Alexa. Deciding what messages are "important" to different people is very hard, and this is a problem that can hopefully be fixed by using ideas similar to using cookies.
Accomplishments that we're proud of
Considered edge cases when implementing functions. Alexa also does what we want :D
What we learned
That connecting everything to the internet is hard. Also that Alexa is harder to use than we thought.
What's next for Alexa Chat Summary
Implement a better natural language processing algorithm and also automatically linking whatsapp chats to a .txt file. Increase the types of chat names since there is no predefined Amazon type for names or common words.
Built With
- alexa
- amazon-web-services
- echo-dot
- node.js
- txt
Log in or sign up for Devpost to join the conversation.