AI MediaKit CLI
An FFmpeg-style command that cuts video locally and sends heavy jobs to the Volcengine cloud: enhancement, ASR, OCR
- 223 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Cloud operations upload your videos and images to Volcengine, and the API key sits in a config on disk or in an environment variable
- A face swap in video is available in the cloud: using it without the consent of the people in frame harms third parties and breaks the law
- Local commands run FFmpeg on files the agent passes in
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/ai-mediakit-cliDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Cloud operations upload your videos and images to Volcengine, and the API key sits in a config on disk or in an environment variable
- A face swap in video is available in the cloud: using it without the consent of the people in frame harms third parties and breaks the law
- Local commands run FFmpeg on files the agent passes in
This is third-party code. Review the repository files before installing.
What it does
AI MediaKit CLI is the official Volcengine command line for video, audio and image processing. The same command either runs FFmpeg on your machine (`--local`: trim, concat, subtitles, audio mixing, audio extraction, metadata) or sends the work to the cloud (`--cloud`): quality enhancement, burned-in subtitle erasing, speech recognition, video OCR, storyline analysis, scene splitting, green-screen and portrait matting. Cloud operations are asynchronous: `mediakit-cli --cloud video enhance-video --video-url <url> --resolution 1080p` returns a task, and `mediakit-cli shared query-task --task-id <id> --poll-complete` waits for the result. Five agent skills come with it: `byted-mediakit-editing`, `byted-mediakit-audio`, `byted-mediakit-image`, `byted-mediakit-video` and `byted-mediakit-shared`. All install with one command, `npx @volcengine/mediakit-cli install -y`, and `mediakit-cli init --mode cloud-first --api-key <key> --yes` sets up the cloud key without prompts.
Who it is for. Developers and content teams who build video-processing pipelines with an agent and are ready to use the Volcengine cloud for AI operations.
Good fit when
- A clip must be trimmed, joined, subtitled or mixed locally without recalling FFmpeg flags
- Low-resolution video must be enhanced to 1080p and you have no capacity for it
- Speech in a clip must become subtitles or text must be read from frames (ASR and OCR)
- Burned-in subtitles must be erased from a video before a new voiceover
- Some pipeline steps run locally and some in the cloud, and you do not want to rewrite scripts when switching between `--local` and `--cloud`
Not a fit when
- You have no Volcengine account: cloud operations need an API key, and without it only local processing works
- FFmpeg 5.1.x with `ffprobe` is missing: local mode relies on them
- Files must not be uploaded to an external cloud: cloud operations send video and images to Volcengine
Example request
Trim in.mp4 from second 3 to second 8 locally, then enhance the result to 1080p in the cloud and wait until it is readyLimitations
The main README is in Chinese, with an English version in `README.en.md`. Installing needs Node.js 18 or newer, and local mode needs FFmpeg 5.1.x with `ffprobe`. Cloud operations are asynchronous and spend resources of your Volcengine account, and the key is taken in the service console. Video translation, narration generation and comic redrawing are marked as coming soon in the README. Availability of the Volcengine console from Russia is not confirmed.
How to disable. Delete the `byted-mediakit-*` skills from the agent skills directory, run `npm uninstall -g @volcengine/mediakit-cli` and erase the stored key from the `mediakit-cli` configuration. Remove the `MEDIAKIT_API_KEY` variable from the environment.
FAQ
How does `--local` differ from `--cloud`?
With `--local` the command runs on your machine through FFmpeg, synchronously and with no key. With `--cloud` it goes to the Volcengine cloud asynchronously, uses an API key and offers AI operations such as quality enhancement and speech recognition.
How do I get the result of a cloud task?
A cloud command returns a task ID. Pass it to `mediakit-cli shared query-task --task-id <id> --poll-complete`, and the CLI keeps polling until the task is done.
How do I configure the CLI without interactive prompts?
Run `mediakit-cli init --mode cloud-first --api-key <key> --yes`. Another option is passing the key through `MEDIAKIT_API_KEY`. The directory for local results defaults to `~/.mediakit/temp`, and `MEDIAKIT_OUTPUT_PATH` changes it.
What does `mediakit-cli doctor` check?
Cloud connectivity and local dependencies (`ffmpeg`, `ffprobe`), and it gives install recommendations. It is the first command to run after installing.
How do I build the CLI from source?
Go 1.22 or newer is needed: `git clone https://github.com/volcengine/mediakit-cli.git`, then `make build`. The skills from the `skills` directory are installed separately with `npx -y skills add ./skills -g -y`.
Discussion
No questions or reviews yet.
Related
CLI toolsA CLI for fal.ai: an agent searches, runs and downloads results from 1200+ image, video and audio models with shell commands.
Local video editing for agents: typed tools over FFmpeg, quality checks and a provenance receipt before you publish
ffmpeg-skill — a skill and MCP server: an agent cuts and joins video, fixes audio, adds captions and exports for TikTok and Reels with local FFmpeg
Video editing for agents: openreelio-cli with JSON output and an MCP server cuts, captions, renders and verifies a clip
A CLI and coding-agent skills assemble rough cuts from your footage and save editable projects for further review
Clipmivo Video
clipmivo-video
A skill for ClipmivoAI: the agent quotes a video task in credits, starts generation, follows it and downloads the result.
README in short
The README lists the strengths, a table of capabilities by area (editing, audio, image AI and video AI), install for humans and for agents, five skills, key-only authorization and the command structure. The `--local` and `--cloud` modes and advanced use are described separately.