gtrk CLI · Open Source (MIT)

Bring intelligent-creation AI into your local agent

One command to install, then any AI agent can call it — describe what you want in Claude / Cursor / Trae and the AI edits for you. The cloud does the heavy lifting; your source files stay local.

Copy this and paste it to your AI — it installs for you

Install the Gitruck gtrk CLI on this machine: run npx @gitruck/cli@latest install and follow the installer. Ask me for my API Key when it is needed (I'll grab it from the website). When it finishes, restart yourself to load the newly installed skills, then tell me in one line how to start editing my first video.

Log in and your API Key gets filled in here automatically, saving a round trip. Paste it to Claude / Cursor / Trae — any AI will do.

After install, use it inside your AI

/gtrk-oralcut edit this talking-head clip, cut the filler and long pauses

🤖 uploaded → editing in cloud…

✅ Client / Jianying / Premiere projects ready, output folder opened

First things first · no chat box in the app

Talk in your agent, watch it in the app

The OpenCut Gitruck Edition desktop app has no built-in AI chat yet. Here is how it works today: you describe what you want in your own AI agent (Claude Code, Cursor, Trae and the like); the agent calls gtrk to produce and edit the project; the app opens that project so you can preview, fine-tune by hand and export the final cut. Every step the agent takes shows up in the app.

1

Talk in your agent

Open Claude Code / Cursor / Trae, call the gtrk skill by name or just describe the edit you want.

2

Watch in the app

Once the project lands, open it in OpenCut Gitruck Edition to preview and fine-tune by hand.

3

Keep talking

Go back to the agent with the next request; the project refreshes in the app. Export from the app when you are happy.

An in-app AI chat is on the roadmap. Development capacity is limited right now, so for this interim period the agent-driven workflow is the one supported path: talk in the agent, watch in the app.

Recommended · Zero barrier

No CLI knowledge needed — just tell the AI

Type a slash command in Claude / Cursor / Trae, say what you want — the Client, Jianying and Premiere projects come back ready.

1

One command installs all

npx @gitruck/cli@latest install — installs the CLI, writes config, and adds the /gtrk-oralcut skill.

2

Restart your AI tool

Restart Claude / Cursor / Trae / WorkBuddy so it picks up the new skill.

3

Type a slash command

Type /gtrk-oralcut and your request — the rest is on the AI.

CLI Capabilities

One CLI, ever-growing creation powers

gtrk is the command-line entry to Gitruck Cloud. From smart oral-cut and beat dispatch to footage search, motion-graphic laying, project conversion and the tool family — same commands, same skills, always growing.

Smart Oral-Cut

Live

Upload raw footage, the cloud trims filler / retakes / long pauses, then pulls back Client + Jianying + Premiere projects and opens them automatically.

Visual Beat Dispatch

Live

Split an edited talking-head piece into beat-level shot assignments — A-roll / motion-graphic / AI reenactment / film B-roll lanes, as a machine-ready worklist.

B-roll Search

Live

Semantic search over your footage library in one sentence; candidate clips are laid into the project for side-by-side picking.

Motion-Graphic Laying

Live

Lint and lay motion-graphic particles onto the project's overlay track — transparent callouts or fullscreen boards in one go.

Project Conversion

Live

Convert a gtrk project into Client / Jianying / CapCut / Premiere / FCPX / OTIO project files — lossless interop.

Single-Point Tool Family

Live

Matting, stabilization, upscaling, frame interpolation, vocal separation, denoising, smart shot detection, one-click MAD… 17 tools on demand.

Film Rendering

Live

Render a project to a finished film locally along its timeline; full cloud rendering on the render cluster is rolling out.

Music Visualizer

Live

Turn a song into a spectrum-visualized film in one command — background / cover / templates / palettes included.

More on the way — video transcription and beyond

Advanced · Command line

Or call it straight from the terminal

Want batch processing, scripting, or precise flags? Same gtrk, same capabilities — you can also run the commands yourself in the terminal.

$ gtrk oralcut footage.mp4 --script script.txt

How the two differ

Agent run (recommended)

You describe the task in plain language and the AI calls gtrk for you. Zero barrier, for most people.

CLI run (advanced)

You type gtrk commands yourself in the terminal. Batchable / scriptable / precise flags, for developers.

Seen enough? Go install it

Going all-in on the full workflow? Head to the download page and grab Full (Client + CLI) — one-click install.