Globrand

Choosing a good brand name is hard enough in your local market. As you grow, what made sense locally could be horrible globally. Globrand lists the pros and cons of a brand name in all languages.

Comment

Inspiration

Lots of people including myself use AI (either directly or through wrapper services) to come up with brand names. But I didn't find any service that would analyze a brand name and tell people what bad meanings that name could have in other languages. (There's a classic example of the Pajero from Mitsubishi and how they had to sell it under the Montero name in spanish-speaking countries because "pajero" means "wanker" there and who would buy a car with that word written on it?!)

What it does

It asks Gemini for negative and positive meanings of names similar to the brand name entered by the user for each language it is aware of, as well as give examples of at most 5 existing similar brand names.

How we built it

I was already working on a bigger side project that would use AI for image editing, so I took some bits from that project and created this small one for the hackathon. I like SvelteKit a lot especially for its simplicity, but didn't get to use it at work yet, so that's why I chose this framework for my side project as well as this one.

Challenges we ran into

The Gemini API wasn't available in Romania, so I had to VPN to the USA to test the project. Something that's still a challenge is having consistent outputs for the same input. Sometimes the responses that I got for the same brand name varied quite a lot. Maybe Gemini is keeping a history of the latest prompts and that influences the current one. There must be a way to turn this history off.

Accomplishments that we're proud of

Prompt engineering so the Gemini response became structured data that can be easily displayed in a table structure. Also proud of the script that automatically generated translation files for the website in 131 languages by using the Google Translate API.

What we learned

  1. The Gemini API is super easy and fun to use, very important in this competitive AI market;
  2. SvelteKit is an awesome web framework (that might surpass React and Vue in a few years);
  3. To stop looking for perfection and just say: if I'll spend too much time polishing it then I'll miss the hackathon deadline :)

What's next for Globrand

If I notice an interest for Globrand following this hackathon, then I'll think about monetizing it for companies that want to check their existing and future brand names.

Built With

Share this project:

Updates