Plan My Trip

PlanMyTrip is a Google chrome extension which can help you plan your trip by just running extension on any Text, Image or Video on internet. The extension uses Google Vision API, Video Intelligence API, Basetrip API and Yatra API in combination.

Project Type

JavaScript, Chrome Extension

Result

We were in top 3 teams out of more than 100 teams but, couldnt win it.

My Contribution

I developed the Google chrome extension with 1 other designer at Yatra Hacktahon.

Plan my trip

Browser Preview

Right clicking any text, image or video would open up the google chrome menu which contained the extension called Plan My Trip and clicking on it would run the extension.

Plan my trip

Text

The text based search was implemented by capturing the text using JavaScript which was then passed to the Basetrip API to get relevant details like Currency, Weather, Language etc. The prices of flights, hotels and packages were calculated by passing the destination and origin(current location) to the Yatra Flight's, Hotel's and Holiday's API. Clicking the Plan Now CTA would take the user to Yatra's search result page.

Plan my trip

Image

The image based search was implemented by capturing image URL using JavaScript and passing it to the Google's Vision API to get the location(destination) which was then passed to the Bastrip API to get relevant details like Currency, Weather, Language etc. The prices of flights, hotels and packages were calculated by passing the destination and origin(current location) to the Yatra Flight's, Hotel's and Holiday's API. Clicking the Plan Now CTA would take the user to Yatra's search result page.

Plan my trip

Video

The video based search was implemented by captuing the URL of video using JavaScript and passing it the Video Intelligence API to get the artifacts(all the possible locations in a video). The process was very slow because Google's Video Intelligence API was processing the complete video to generate artifacts and more than 1 artifact was returned by the API so, basically the results were not precise. This approach failed because Video Intelligence API was new and it was not trained properly. This approach might work in future when the API's will become more mature.