Universal Translator

Inspiration

Videos on facebook advertise 'magic' translating solutions, usually featuring expensive hardware. We wanted to leverage the software that allows for this function already, in a minimal design, that is free and open source.

What it does

Takes 2 inputs, the language of the users spoken word and the language to which they wish to translate it. Originally, we used App Inventors built in speech to text component but we moved on to a custom extension made by Taifun, which allows more language inputs than just english (both use google voice to text). Then using another built in component, yandex translator, the app translates the resulting text from the tts and returns the desired translation and prints it on the screen. Afterwards, an api, voicerss, is used to take the translation and convert it to spoken word, which is then played back to the user.

How I built it

MIT App Inventor and a lot of googling/drag and dropping. Really though, most of the software was given, like the voice to text, translation and text to voice. We only configured those softwares to work with each other.

Challenges I ran into

Configuring the api for text to speech was the hardest part, because it was the only component not made with App Inventor in mind. We also wanted to implement a feature that would use the devices camera to take aa picture of foreign text and using an ocr api, yandex translator, and tts, scrape the text from the image, translate it and dictate it. We had to scrap this feature (although remnants exist in the app) because we could not get the image to upload to the api or get the app to parse the JSON.

Accomplishments that I'm proud of

I had never encountered the concepts of post and get before, so being able to get the tts api working was an accomplishment for me. Getting at least a response for the ocr feature was an accomplishment too because we had no idea where to start with that function.

What I learned

POST and GET protocol.

How to leverage an API.

how to speak some choice phrases in other languages.

What's next for Universal Translator

We seek to improve the layout, the language options are not intuitive, the spoken translation is very soft, etc. Still want to implement the ocr feature. We will hold strongly onto our original values, minimalism and open source software.

Built With

• mit-app-inventor
• taifun-tools
• voicerss