Horus

• 

  BI for WordPress plugin devs

Inspiration

Selling premium WordPress plugins is difficult given the chaotic nature of the open source software. We have been website developers and tried selling plugins too but did not have a clear idea if the plugins we have developed for some clients were even "sellable" to more clients.

Are there enough clients using this precise "recipe of plugins" so that this new plugin we developed for the client can potentially be productised and sold to more websites?

We did not find a proper answer to that question without doing some data analysis ourselves so we want to build a tool that can do it for other plugin developers.

What it does

Provides real metadata and insights on the plugins stack and traffic data from WordPress websites. Plugin developers can get business intelligence to sell more. They can expand onto new markets and figure out what features will be needed by their existing and potential clients.

How we built it

Chatbot using LLAMA, Mongo DB and the vector search capabilities of Mongo. We also scraped data from a large amount of WordPress websites as we needed their "recipe".

Challenges we ran into

The AI services of the ThetaEdgeCloud was unavailable very often. We also need to keep scraping more data and re-scrape data from websites we already have since the data will become stale after some time.

Accomplishments that we're proud of

We managed to build a useful chatbot in limited time.

What we learned

Vector Search principles and how to use Langchain with LLAMA 3.

What's next for Horus

We should build a plugin that website owners can install on their websites and they report which recipe they are using. A specific data sharing agreement on which they are rewarded should be worked out.

Built With

• langchain
• llama
• python
• typescript