Visionary

Inspiration

We were inspired by how much great content never gets shared. This isnt because it isn’t good, but because finding it is painful. Raw footage is chaotic, and traditional editors force creators to manually hunt for moments they feel are important. We wanted to flip that model and build something that discovers moments for you, the same way a human editor with great taste would.

What it does

Visianary runs quietly in the background while you upload or record video. It detects rare, emotional, or unexpected moments and automatically turns them into polished, share-ready clips with smart cuts, captions, and voiceover. There are no prompts, timelines, or configuration—just instant discovery.

How we built it

We used Daytona AI to orchestrate background agents that analyze video streams for visual, audio, and temporal signals. ElevenLabs powers expressive, cinematic voiceovers that adapt to the tone of each moment. CodeRabbit helped us rapidly generate and iterate on the editing pipeline, event-detection logic, and frontend interactions. The frontend is built with Next.js, Tailwind, and Framer Motion to deliver a motion-first, consumer-grade experience.

Challenges we ran into

The biggest challenge was avoiding a “generic AI tool” feel. Most UI patterns push users toward prompts and controls, but we intentionally removed them. Designing interactions that felt intuitive without explicit buttons or settings required a lot of iteration. On the technical side, synchronizing video playback, moment detection, and real-time animations without breaking immersion was extremely difficult as well, and we had a lot of bugs. Sentry helped tons with this.

Accomplishments that we're proud of

Built a fully working end-to-end demo in hackathon time
Created an editing experience with zero prompts or timelines
Achieved seamless moment discovery that feels automatic and inevitable
Delivered a cinematic UI that prioritizes taste and motion over configuration

What we learned

We learned that the hardest part of building AI products isn’t the models—it’s deciding what not to expose to the user. Great AI feels invisible. We also learned how powerful background agents can be when paired with strong product intuition and thoughtful UI design.

What's next for Visionary

Next, we want to expand real-time detection, support live streams, and personalize moment discovery based on creator style. Longer-term, we see Visianary becoming the default way moments are found—not just edited—across video, streams, meetings, and beyond.

Video: https://www.loom.com/share/18d0803e37724213aa94cc2c612275b5