Keyboard-driven, Vim-style video editor for terminal workflows
Problem
Video editing remains locked to mouse-heavy timeline GUIs. Developers and text-centric users who live in keyboards (Vim/Emacs) have no editor that matches their workflow: one commenter resorted to hacking Emacs with subtitle-based timespans to do basic audio/video cutting. Power users want modal, keyboard-first editing for video the way Vim did for text.
Opportunity
A modal, keyboard-driven video editor (terminal or minimal GUI) with Vim-style verbs for cut/trim/rearrange, scriptable editing, and reproducible edit definitions, targeting the developer/creator niche.
Market analysis
The itch is real for a small crowd, but video-use from the browser-use organization just validated and occupied the text-first, no-timeline editing approach with large open-source traction. What remains for a Vim-style editor is the human precision-editing niche, which is passionate, tiny, and expects free software.
Market · Developers and text-centric creators doing rough-cut editing of screencasts, podcasts and talks; small but vocal niche.
Pricing · OSS expectations dominate this audience; comparable text/EDL-based tools are free, so any paid version would need a pro-creator angle to justify a subscription.
Pros
- + The Emacs-subtitle-hack workaround proves motivated users are jury-rigging solutions today.
- + Reproducible, text-based edit definitions compose naturally with git, scripting and automation.
- + Modal verb grammar for media is an unexplored UX space with a ready-made mental model (Vim).
Cons
- − video-use (browser-use org) already owns the 'editing as text, no timeline' mindshare with wide adoption.
- − Preview performance and frame-accurate seeking in a terminal or minimal GUI are genuinely hard.
- − Market is small, opinionated and expects OSS; passion-project economics, not revenue.
Existing / similar tools
- → video-use ↗
- → cut-ai ↗
- → Remotion ↗
Source
Hacker News (Ask HN)
It is worth being clear-eyed: video-use’s success is both the validation and the occupation of this idea. It proved that treating a video as a transcript plus an ffmpeg edit-decision list beats dragging clips on a timeline, but it did it for AI agents, not for humans. The remaining space is the human counterpart: a modal editor where dw deletes a word of speech, 3dd cuts three sentences, and the edit is a diffable text file. That is a beautiful, buildable thing and a genuinely better rough-cut experience for keyboard people, but it is a Vim, not a VS Code: beloved by few, used for real work by fewer, and monetized by essentially nobody. Build it for love or not at all.