-
-
-
The login page for Hermes.
-
The home dashboard. Here you can find your most recent chats, and start a new one.
-
When starting a new chat, the system will prompt your conversation partner to speak so it can detect their language.
-
During your chat, it will find previous chats you've had with your partner, and translate both of your communications to the other.
-
During your conversation, you can take notes of what they say for personal reflection.
-
When you think you are done, Hermes will make sure this is the end of the conversation before ending the chat.
-
Once the chat is complete, you can review what you discussed through AI summaries, as well as sharing the conversation to those need it.
❓Problem & Opportunity
Do you have a relative whose primary language isn’t English? Access to communication, care, and travel is full of blockers that without personal assistance, can inhibit their daily lives. Recognizing a need in the healthcare, tourism, and foreign language journalism market, we wanted to create a tool that could break through that language barrier and create an easier experience for service providers.
✅How it helps
Hermes creates a conversational translation experience that removes the broken and frustrating process that Google Translate presently offers. By keeping a conversational log of what's been said by both parties, we're able to simplify the recording flow into a constant ongoing process, rather than a one sentence at a time model. We’ve removed slow and clunky human intervention to the translation space.
🛠How we built it
Thanks to the generous sponsors of Hack the North 2024, accessing the technologies needed to launch Hermes was a breeze. We used React, Google Cloud APIs, Convex Backend, Speech Recognition, Firebase, Cohere API.
💪Challenges
With the short time constraint, as well as financial constraints, we had to pull back on features we wanted for our MVP. One challenge that arose was understanding which APIs we had available had the best use case for our project. The biggest and probably most relatable challenge, was matching up the variety of skill sets up to work in a consistent environment. Understanding each others proficiencies in JavaScript, TypeScript, React, and CSS was crucial in ensuring we would have an equitable division of labour to make sure we'd be able to deliver on time.
😤What we’re proud of
First and foremost, we are proud to be given the chance to solve a problem that can impact many Canadians, and people worldwide. With the advent of AI and LLMs paving the way for process optimization and innovation, we are proud of our ability to learn and utilize resources like Convex, Cohere, OpenAI, and countless others in such a short period of time.
🔜What’s next
We were building Hermes with the Canadian market in mind, but the app itself delivers on an experience that can have a global impact. These days borders are becoming more of a social construct than a literal one and access to communication across communities is becoming more and more vital. Even within countries with multiple languages across large populations, the opportunities are endless. Scaling this app into a robust and mature product would service a larger demographic and allow anyone to communicate with their community. With that, we have three main goals in mind to grow Hermes:
1) Add more languages. For our MVP we limited ourselves to only a couple of languages to keep the systems simple for our launch. We want no gaps in our service, and expanding to add more languages is crucial.
2) Expand into more markets like education and athletics.
3) Move away from English as the primary language. Only 15% of the world speaks English. That's 85% of potential customers left unserved and unable to communicate. Hermes will bridge that gap by allowing the app to be provided in multiple languages, as well as allowing translations from language to language,
Log in or sign up for Devpost to join the conversation.