Vibe Video

Inspiration

Video editing is stuck in the past. Tools like Premiere and Final Cut are powerful but painfully manual — you spend hours dragging clips around a timeline for what should be a creative process. Meanwhile, coding got "vibe coding" with Cursor. We asked: why can't video editing work the same way? Throw in your raw footage, get a real edit back, then iterate on it conversationally.

What it does

Vibe Video is an AI-powered video editor that turns raw clips into a fully edited video — then lets you refine it like you'd refine code in Cursor. Drop in your footage, and the AI assembles a coherent edit with cuts, ordering, and flow. From there, you can edit the transcript directly to change the video, give natural language instructions ("make the intro punchier," "cut everything after the second interview clip"), and watch your changes apply in real time. It's vibe editing.

How we built it

We built the frontend as a web app with a transcript-based editing interface that mirrors the Cursor-style inline editing experience. On the backend, we use AI to analyze and transcribe the uploaded clips, then intelligently sequence and cut them into a cohesive edit. The transcript serves as the single source of truth — edits to the text propagate back to the video timeline, so changing words literally changes the video.

Challenges we ran into

Video processing is slow. Getting the feedback loop tight enough that editing felt conversational rather than batch-job-and-wait was a constant battle. Syncing transcript edits back to precise video cuts without weird jumps or artifacts was also trickier than expected — off-by-a-frame errors add up fast.

Accomplishments that we're proud of

The core loop actually feels good. You can go from a pile of raw clips to a watchable edit in minutes, and the transcript-driven editing is genuinely intuitive. The fact that it works end-to-end in a weekend is something we're proud of.

What we learned

Working with video programmatically is a whole different beast compared to text or images. We gained a much deeper appreciation for how hard real-time media manipulation is — and how much opportunity there is to make it better with AI.

What's next for Vibe Video

Multi-track editing, AI-suggested b-roll and transitions, and support for longer-form content like podcasts and vlogs. We also want to add a collaborative mode where multiple people can vibe-edit the same project. The dream is to make video editing as fast as thought.

Built With

• fastapi