When (like me) english is not your native language, it can be hard to remember exactly what was said during a meeting. I also know that if I have a paper in front of me and a pencil in my hand, I will probably (after an hour or...) start drawing and stop focusing on the meeting. Browsers and phones can now understand us talking to them. So we took everything, added some Redbull and got that idea...
What it does
The website allows you to log, create a meeting room and invite people to it. Once you are done with inviting everyone, you can push the start button and it will start recording the speech of every invited users. You can pause the meeting, or stop the meeting. Once stopped, you will be able to see the text recording again or download a text version of it.
How we built it
Using meteor and the Web Speech API (Beta).
Challenges we ran into
The speech API is made for a special situation: "The web browser needs to decode the speech of ONE user and turn it into text", it's not made for a conversation, so the challenge was to make it work with multiple people.
Accomplishments that we're proud of
We reached all of our goals, and made a functional website.
What we learned
I (Arnaud) learned Meteor! And I also learned some stuff about webkit audio tools and google app engine.
What's next for Transcripter
A lot! Improve the conversation system, add a 'friend list', correct some bugs, edit, security... And also, recording the audio and let the user find something in the audio file by clicking on the text.