• During the coronavirus crisis, every business is affected. But big companies have budgets and manpower to do social media marketing, moving sales of products or services online, resulting in incomes that are not decreasing or may increase even more

  • But a small business or a family business doesn't have enough people, has a limited budget, doesn't have expertise in social media marketing and doesn't have the ability to move sales or services from offline to online, resulting in many small businesses have decreased revenue or had to shut down.

What it does

  • Thrive therefore developed an application that allows users to upload images or videos of a product or service with the name and price of the product and record audio that briefly describes those products.

  • Our system separates images with deep neural networks, arranges images and videos into categories, and provides auto-generated descriptions for those images or videos. Users will get a preview of the post with pictures and product details that are automatically generated.

  • When the user presses the confirm button, the application will post to all social media platforms that the customer has defined. The application will automatically post again every week with an updated description. Users will have to press confirm to repost or edit the description to control the quality of the post. Users can schedule a time for reposting themselves.

  • The system will summarize every post, how many likes, how many shares, and which photos are popular. What products should be posted during which days and times?

How I built it

  • Our team builds web pages with basic HTML CSS javascript with Bulma UI and uses Python flask as a backend.
  • We use google speech-to-text API to parse the frontend speech into text and then generate the words with transformers library using PyTorch.
  • Run the pre-trained model "gpt2" to create words from the context in which the user speaks and returns to create a preview page in the frontend.

Challenges I ran into

  • It's hard to think of a business model for this application.
  • It must be carefully planned because the system is more complex than we thought.

Accomplishments that I'm proud of

  • Create a minimum viable product within 24 hours, along with a landing page, a demo video, and product documentation.

What I learned

  • We learn that when we create an application we must also think about who it was created for, what its benefits are, and how it is used.

What's next for Thrive

  • We will start using this application immediately after the end of World Hackathon Day with our restaurant in Bangkok.
  • We will collect feedback to develop applications to best meet the needs of users.

Built With

Share this project: