GenPost

GenPost helps medium and small businesses growing their presence on social media platforms. GenPost uses Square items information to generate catchy posts by leveraging the power of Google AI.

Comment

GENPOST

Create social media posts easily using Square and Google Generative AI. Please try it out at https://capable-meerkat-e8ac3d.netlify.app/

Inspiration

We realized that small and medium businesses may have little time to curate social-media platforms because they may lack resources (time and/or money) to invest in this activity, even though it is a fundamental part of all businesses nowadays. We also realized that generative AI is really powerful when it comes to creating content, which is a key skill in marketing campaigns. That's why we decided to explore how generative AI can solve the aforementioned problem and that's why we created GenPost: a simple but effective platform that unites that huge API infrastructure of Square with the power of Google AI

What it does

GenPost generate social-media posts by combining the following inputs:

  • Square items
  • User prompt
  • Business description
  • Business type

The idea is that the user can quickly and easily select items that they want to advertise with a post and the user can edit the prompt to tweak the post the way they like.

How we built it

The frontend is a static web-site written in React. It just interfaces with a private API implemented in the backend. The backend runs with Python and Flask with MongoDB as database to store the user information. The authentication for GenPost works using Square OAuth API, so merchants do not need another account that it is easily forgotten. GenPost combines the generative AI features of PALM API and Square Catalog API to generate the social-media post. The items selected by the users are fetched from Square and a description is generated and fed to the PALM API model.

Challenges we ran into

We wanted to add OAuth authentication for social platforms to simplify the posting operation, but we had to drop this feature for the time being because some APIs are very contrived. We also had some problem during the release phase with clashing dependencies of Python packages. We also found out that react dynamic routing does not work properly in all the deployment environments. Using the image API of Google AI was difficult and with very little documentation, so we had to drop the feature of using images uploaded on Square to give more information to the generative model.

Accomplishments that we're proud of

Deployed a Flask server for the first time! We are also happy that we managed to develop a platform like we envisioned at the beginning.

What we learned

We learned a lot about which Google AI APIs are available and we think we are going to explore this further. We also have grown knowledge of React and Flask.

What's next for GenPost

  • Integration with social-media platforms
  • Usage of images stored with Square with Google AI APIs to increase the quality of posts
  • Get feedback from users to increase quality of posts

Built With

Share this project:

Updates