Back in January, we had the opportunity to support Hack Cambridge—a student hackathon sponsored by the University of Cambridge. We ran a challenge for participants to use Deepgram in some way to create a fun and unique hackathon project. Developer outreach and support is a big focus for the Deepgram team, and supporting these students in building cutting-edge projects using our API is a passion for the whole company—we're always pumped to see folks attending events and running hackathons. I don't think it's any exaggeration to say that we were absolutely blown away by the quality of the projects that incorporated Deepgram and used speech-to-text to create something new. Over the last few months, Kevin Lewis, a Senior Developer Advocate here at Deepgram, has been publishing descriptions of some of the projects that used Deepgram. Read on to learn more about the incredibly creative projects that came out of Hack Cambridge 2022.
Deepgram Hack Cambridge Projects
Here's a list of eight of the best projects from the hackathon. Each one is linked to Kevin's full post about the project on our developer blog if you'd like to read more or take a look at the code for the projects.
- Add Live Speech Bubbles to Youtube Videos with AutoBubble - The winning project of our challenge was by AutoBubble, which uses Deepgram not just to add captions to video, but to add those captions as speech bubbles into the video-super cool!
- Voice Control Your Browser with Stëmm - We're a big fan of saying that voice is the next interface-and as speech-to-text tools get better and better, voice interfaces will continue to become easier to use and more powerful. The Stëmm team took advantage of Deepgram to build a system that would allow for voice control of Google Chrome, doing everything from opening new tabs to searching Google.
- Create Comic Books from Videos with yack! - Using a combination of computer vision and Deepgram, the yack! team built a tool that takes a video as an input, chooses single frames of the video to create a comic book, and then uses Deepgram to transcribe the audio and create speech bubbles to go along with the images. Try it yourself at yack.ml!
- Creating Contextual Video Overlays with TomScottPlus - Ever wanted to know more about a topic mentioned in a YouTube video? TomScottPlus has your back. Using a transcript generated from a YouTube video, the tool overlays information from Wikipedia that's relevant to what's happening on screen-a sort of modern Pop-Up Video.
- Sharpen Your Foreign Language Skills With Triolingo's Chatbot - There are lots of language-learning apps available, but few provide a real chance to practice speaking skills, never mind actual conversations. But by using Deepgram's Python SDK and a chatbot powered by GPT-3, Triolingo gives users a chance to practice their oral language skills, even if they don't have a conversation partner.
- Use Your Voice to Draw with ARTiculate - If you've never thought about how you might draw verbally, you're certainly not alone. But for people who can't use traditional input devices, the ability to draw with their voice has the potential to unlock their creative potential. This project was created to provide a way for people to draw who can't use traditional input devices, using Deepgram to transcribe their commands.
- Collaborative Augmented Reality Note-Taking with Airnote - The Airnote project combines augmented reality, note-taking, and Deepgram's speech-to-text API to create a collaborative experience to transcribe notes.
- Practice Spelling Bees with Spelling Hero - If you get pumped about spelling bees and dream of competing someday, you've now got a tool to help you practice and get ready for your next bee. Spelling Hero lets you set up drills with varying levels of difficulty, and uses Deepgram to transcribe your oral answer for the full bee experience.