fastCutVid is a lightweight native video-cutting tool designed with automated editing by AI agents in mind. You can trim and reorder imported footage on a timeline and export it as MP4. Its defining feature is that edits are saved as a lightweight JSON timeline. Humans can review that same file visually in the desktop app, while AI agents can use it from the CLI for validation and rendering. It ships with skills for Claude Code / Codex, requires no cloud uploads or API keys, and runs locally alongside FFmpeg / FFprobe. It is an open-source (MIT) project by the independent developer modecir, and its first release was published on September 5, 2026.
Key Features
- A native GUI focused on timeline editing: A desktop app built with Rust (egui) that lets you assemble a sequence by trimming and reordering footage. Effects and title compositing are deliberately out of scope, so the design concentrates on “cut, arrange, export.”
- JSON timeline (.fastcut format): Cuts are saved as a portable JSON file. It is easy for both humans and AI agents to read and write, so edit proposals can be handed off as a single file.
- CLI renderer: Commands such as
--validate,--render, and--versionlet you validate a timeline and export MP4 from the terminal. An agent can go all the way to rendering without opening the GUI. - Bundled skills for AI agents: The repository includes
.agents/skillsandAGENTS.mdfor Claude Code / OpenAI Codex, which help agents build timelines. Note that no MCP server is officially provided. - Fully local, with FFmpeg integration: It uses ffprobe for media analysis and ffmpeg for export. Footage never goes to the cloud, and no API key is needed (FFmpeg must be installed separately).
- Cross-platform: Builds for macOS (Apple Silicon / Intel), Windows, and Linux are distributed via GitHub Releases.
Pricing
| Plan | Price | What’s included |
|---|---|---|
| Free (open source) | Free | No feature limits. Released under the MIT license. Use the prebuilt binaries or build from source |
Pricing is as of September 2026. For the latest information, see the official website and the GitHub repository.
There are no paid plans or pricing page; it is distributed free of charge as MIT-licensed open-source software.
Pros & Cons
✅ Pros
- You can edit without uploading footage to an external service, which makes it easier to handle material you don’t want to leave your organization.
- A human can check an edit proposal made by an AI agent on the same timeline before export, so unintended cuts can be caught before the final render.
- Because the result remains as a JSON file, it is easy to review diffs, put edits under version control, and re-render in another environment.
- Thanks to the MIT license, it is free to use, and you are free to modify it for internal use or build it yourself.
⚠️ Cons
- It does not support effects, transitions, color adjustment, or captions, so you will need separate editing software for finishing.
- FFmpeg / FFprobe must be installed separately, which adds a setup step.
- It is an independent project released only in September 2026, so there are still few use cases and little information available.
- On Windows / Linux, preview playback is silent (as of v0.1.4). Audio is included only in the exported MP4.
- Early macOS builds are not notarized by Apple, and the Windows / Linux packages are unsigned, so the OS may show warnings at launch.
Comparison with Similar Services
| Service | Difference from fastCutVid |
|---|---|
| LosslessCut | An open-source GUI video trimmer. Strong at lossless cuts, but has no JSON timeline aimed at AI agents |
| Shotcut | Full-featured open-source video editing software. Supports effects and transitions, but is not specialized for lightweight cutting |
| DaVinci Resolve | Professional, high-end editing software with a free version. Not designed around a CLI renderer or agent integration |
| ffmpeg (CLI) | The standard command-line tool for video processing. Powerful, but has no timeline GUI or way for humans to check cuts visually |
Who Is It For
- People who want Claude Code or Codex to cut unnecessary parts from interview or course videos, then check the result visually before exporting
- People who mostly do light editing ─ quickly trimming and reordering several clips into a single MP4
- People who want to edit internal videos or personal footage on their own computer without uploading them to the cloud
- Engineers who can use ffmpeg commands but want a screen to check cut points
Summary
The challenge when handing video editing to AI is how a human verifies the cuts an agent decided on. fastCutVid tries to answer that by consolidating edits into a shared file ─ the JSON timeline ─ so the agent’s work and the human’s review sit on the same foundation. Its features are limited to trimming and reordering, and because it has only just been released there is still little information about it, but for anyone who wants to try an agent-assisted video editing workflow, it is an option you can start with for free on your own machine.