Reverse Image Search in Twitter

Twitter as we know it is a great platform to share, learn and have fun observing the left and right fight :D. While the platform itself is evolving and improving constantly, there are some features that I felt are needed but are not available. That’s where Twitter's APIs came handy. They provide opportunities to try new ideas and add features to twitter in the form of Bots. I found couple of features that I thought could add good value to Twitter:

There are countless Images shared in Twitter without any more information, without context, without mentioning the source or even worse with wrong context or wrong source. When some accounts share images like this, there are replies asking for What, Where, Who etc. I saw this happening in too many places to ignore. A Reverse Image Bot for Tweeted images made a lot of sense. So, I got into action and @SearchThisImage bot was born. (How I did it is for another day / blog).

After releasing the bot, I started to use it myself. I used it mainly in the Archillect and other accounts where image is shared mostly without context. Users saw it and started following the bot. Currently it has 900+ followers. I see it is being used by tweeples mainly in the below two cases: 

  1. To Get more Info This is the case I describe above where there is no or very less information shared in the image tweet itself. Users use the Bot and go to the result web page which shows them some more details. Example1, Example2
  2. To Validate the Image Context. Many tweets combine text message with a related image to convey the point better. As Twitter says, Tweets with Images have better reach than tweets without. But the images are sometimes very misleading and provide wrong information. Reverse Searching the image leads them to a website that provides the correct context. Example1, Example2
  3. To Credit the actual Creator. Digital Arts like Photographs and Paintings are widely shared in Twitter. We could see some users claim to have created the art explicitly or implicitly; but they haven't actually. It is easy to claim ownership to an unknown audience. This Bot has been useful in validating those claims. Example

Future Plan:

Though the current result provides useful information, many times providing just one link is very limiting. I've seen more information is avialable in the subsequent links (Google provides 10 links in the API request). The plan is to publish rest of the links in this webpage It will be dynamic page - one for each image.