The Vivideo CLI brings AI video generation to the terminal. Run one command to generate text-to-video, image-to-video, avatar or voice clips across 30+ models — scriptable and pipeable, so you can batch-generate, wire it into CI, or generate video from a shell script without writing API code.
Trusted by individuals at
How the AI video CLI works
Install the CLI
npm install -g @vivideo/cli, then authenticate with vivideo configure. Runs anywhere Node 18+ does — macOS, Linux and Windows.
Run a generate command
Pass a prompt (or an image URL) and a model flag — vivideo create submits the job and --wait tracks it to the finished video.
Batch or pipe it
Loop over a file of prompts, pipe output between commands, or wire it into a CI step for automated video.
Get the video URL
The CLI prints the hosted video URL when the render completes — pipe it to curl in a script, or read it as structured JSON with --json.
What the AI video CLI gives you
Every model and generation type, scriptable from the terminal.
| Capability | What it does |
|---|---|
| 30+ models from the terminal | Generate with Sora, Veo, Kling and more via a model flag. |
| Scriptable & pipeable | Structured --json output and meaningful exit codes — loop over prompts, pipe between commands, branch on results. |
| Batch generation | Generate hundreds of clips from a file of prompts in one run. |
| CI-friendly | Wire video generation into pipelines and automated workflows. |
| Same power as the API | Text-to-video, image-to-video, avatars and voices, all from the shell. |
Developer resources
Install, reference and source — start scripting video.
CLI reference
Every command, flag, exit code and JSON shape, with examples.
@vivideo/cli on npm
One npm install gives you the vivideo command everywhere.
API quickstart
The REST API behind the CLI — same key, same models.
Agent skills
Let AI agents drive the same video workflows safely.
Open source on GitHub
MIT-licensed toolchain — CLI, SDK, MCP server and skills.
Get an API key
Create and manage keys in your Vivideo account settings.
How the AI video CLI works, explained
Not every video workflow belongs in a browser. When you need to generate hundreds of clips, wire video into a build step, or script generation alongside other tools, a command-line tool beats a UI. The Vivideo CLI brings AI video generation to the terminal: one command generates text-to-video, image-to-video, avatar or voice clips across 30+ models.
Because it's a CLI, it composes. Loop it over a file of prompts to batch-generate variants, pipe its output into the next command, or drop it into a shell script or CI pipeline so video gets made automatically. Everything the API can do is a command you can script — with structured --json output, meaningful exit codes and automatic idempotency keys so retries never double-charge.
Model choice is a flag. Select Sora, Veo, Kling or any of 30+ engines per command, so a batch job can route by cost or style, and a pipeline can fall back to a cheaper model under load. The CLI submits the job, tracks it with a safe bounded wait, and prints the hosted video URL your script can download or pass along.
It's the same generation power as the AI video API and MCP server, shaped for the terminal and automation. The CLI is open source (MIT) on GitHub, installs from npm as @vivideo/cli, and is documented at developers.vivideo.ai — install it, authenticate with your key, and generate video from the command line: scriptable, pipeable, and ready for batch and CI.
AI video CLI — FAQ
What is the AI video CLI?
It's a command-line tool for generating video. Run text-to-video, image-to-video, avatars and dubbing across 30+ models from your terminal — scriptable and pipeable, no API code needed.
How do I install it?
npm install -g @vivideo/cli, then authenticate with vivideo configure or the VIVIDEO_API_KEY environment variable. It runs anywhere Node 18+ does — macOS, Linux and Windows.
Can I batch-generate videos?
Yes — loop the CLI over a file of prompts to generate hundreds of clips in one run, ideal for variant testing or bulk production.
Can I use it in CI?
Yes — --json prints structured output, exit codes distinguish auth, credit and rate-limit failures, and automatic idempotency keys make retries safe, so pipelines never double-charge.
Which models can I use?
30+ — Sora, Veo, Kling and more — selected with a model flag, so you can script the right engine per job.
Does it support image-to-video and avatars?
Yes — text-to-video, image-to-video (pass an image URL), avatar and voice videos are all commands, alongside credit estimates, status checks and account queries.
How is it different from the API?
The CLI wraps the API for the terminal — same models and capabilities, but as commands you can script and pipe, no code required.
Is the CLI open source?
Yes — MIT-licensed on GitHub as part of the Vivideo toolchain and published on npm as @vivideo/cli, with the full command reference at developers.vivideo.ai.
AI Video From the Command Line
The Vivideo CLI generates video from your terminal — text-to-video, image-to-video, avatar and voice videos across 30+ models. The answer to "how do I generate video from a script" without writing API code.
One command generates a clip; a loop generates a batch.
Scriptable, Pipeable, CI-Ready
Loop over a file of prompts, pipe output between commands, or wire the CLI into a build step — video generation composes with everything else in your shell.
Batch hundreds of clips, or make video part of your pipeline.
The API's Power, in the Terminal
Every model and generation type the API offers, exposed as commands you can script — select the model with a flag and route each job however you like.
Same models, same capabilities, no integration code.
Everything the AI Video CLI Includes
One command-line tool for every model and generation type.
Who Uses the AI Video CLI?
Developers and teams automating video:
One Free Studio Behind Every Tool
Each utility here is a doorway into the full Vivideo studio — create, edit and localize video in the same place.
Explore More Developer Tools
See the AI Video CLI in Action
Watch one command generate a video — then install the CLI on Vivideo.
