...
 
Commits (3)
# Notia: No more note taking during calls. Notia records your calls, and transcripts them in specific tasks.
# Notia: No more note taking during calls. Notia records your calls, and transcripts them into specific tasks.
‘A pencil is a bicycle for the mind’ - Albert Einstein
......@@ -10,16 +10,38 @@ Notia solves this problem by recording your calls, and transcribing them in to s
This solution was made with NodeJS (back-end), Android Native (recording app), VueJS (progressive web app front-end), Google Dialogflow (AI & NLP), much love and without sleep. With the working prototype you can convert your conversations into text (NLP). From the text, actions point are determined by our AI algorithm made with Dialogflow.
#### Table of Contents
* [>> Concept](#concept)
* [>> Technologies used](#technologies-used)
* [>> Design](#design)
* [>> Concept & technology](#concept)
* [>> How it's made](#how-its-made)
* [>> Recognized tasks](#recognized-tasks)
* [>> Technologies](#technologies)
* [>> Designs](#designs)
![Design impression](Design/impression-5.png)
## Concept
## Technologies used
Notia is a automatic note taker for your calls. It records your calls, and transcripts them into specific tasks. This makes phone calls far more conveinient, especially when you don't have your both hands available (e.g. in the car).
#### How it's made
We wanted to use the Nokia API to record the calls, however this was not available yet (but will be soon they told us 😄). In accordance with the Nokia mentors we decided to hack it ourselves using a Native Android part. This was harder then expected, because we didn't have much experience with Android and there is no clear documentation available. But after long hours, we managed to hack it!
The Android part streams the calls to the NodeJS back-end. The back-end processes the speech audio into text using Dialogflow (Natural Language Processing algorithm from Google). That way we convert the speech to text. Dialogflow is also used to determine the specific tasks.
#### Recognized tasks
Therefore we set-up an AI chatbot and trained it to recognize the following tasks:
* Email / Mail
* Meeting / Appointment
* Call / Hangouts / Skype
* Other tasks are marked as 'Other'
The AI recognizes synomyns as well and learns new ones by itself. Example sentences that you can use:
* "Hey Anna, how are you? I will __email__ you the documents __this afternoon__"
* "Hi John, let's __have a coffee__, I propose __tomorrow__ at __5pm__"
* "I will schedule the __call__ with Jonathan __next week__"
* "Thanks for the reminder Sasha, we will __send__ the package __tomorrow__."
#### Technologies
We used some of the newest technologies available to develop and design this solution.
* NodeJS for the back-end
......@@ -31,7 +53,7 @@ We used some of the newest technologies available to develop and design this sol
## Designs
We put a lot of effort in the designs. The UX design was made with paper prototypes, the UI desings in Sketch. [Click here all the designed screens and source files](Design/)
We put a lot of effort in the designs. The UX design was made with paper prototypes, the UI desings in Sketch. [Click here for all the designed screens and source files](Design/)
![Design impression](Design/impression-1.jpg)
![Design impression](Design/impression-2.jpg)
......