VibeStudio

VibeStudio MCP Server is a headless, zero-runtime video editing engine built with Bash scripts and FFmpeg. It exposes powerful media editing actions as plain shell tools you can call from any MCP-enabled assistant, enabling fast, traceable, tool-based workflows without extra runtimes.

How to use

You interact with VibeStudio through an MCP client by calling its available tools to perform common media tasks. Start a session, list the tools you can call, and then invoke a tool with the required inputs. Each tool returns structured results describing inputs, outputs, and any generated artifacts. Use the outputs to chain multiple steps, for example extracting audio from a video, then resizing the video, and finally exporting a thumbnail for previews.

How to install

Prerequisites you must have before installing VibeStudio MCP Server are a Bash shell (version 4.0 or newer), jq for JSON processing, and FFmpeg with common codecs.

# Prerequisites

# 1) Install jq
# macOS
brew install jq

# Ubuntu/Debian
sudo apt-get update
sudo apt-get install jq

# CentOS/RHEL
sudo yum install jq

# 2) Install FFmpeg
# macOS
brew install ffmpeg

# Ubuntu/Debian
sudo apt-get update
sudo apt-get install ffmpeg

# CentOS/RHEL
sudo yum install ffmpeg

# 3) Bash is required (version 4.0+ usually comes by default on modern systems)

Configuration and startup notes

VibeStudio is designed to be run locally as an MCP server that your assistant can call. The main MCP entry for local use is a standard stdio server script that you execute directly in your environment. You configure the client to start the server process and then point your MCP client to the local stdio interface. The server exposes a stable set of tools for video and audio editing, as well as media analysis.

{
  "type": "stdio",
  "name": "vibestudio",
  "command": "/absolute/path/to/vibestudio-mcp/ffmpegserver.sh",
  "args": [],
  "env": {}
}

Available tools you can call

The server implements a broad set of media operations. You can convert formats, trim, merge, resize, add subtitles, extract frames, create videos from images, apply filters, generate thumbnails, extract and convert audio, and normalize volume. Each tool accepts an input file and outputs the desired result, along with additional configuration to tailor the operation to your needs.

Examples of practical workflows

Common tasks you can perform in sequence include converting a video format, resizing to target dimensions, extracting audio for separate processing, and creating a preview thumbnail for the project. You can also burn subtitles into a video, extract high‑quality frames for thumbnails or analysis, and generate a new video from a sequence of images.

Troubleshooting and tips

If you encounter issues, verify dependencies are installed and executable permissions are correct. Check that your input files are accessible from the script location. If the JSON input or tool invocation fails, consult the logs to identify which tool encountered a problem and inspect the inputs you provided.

Security and maintenance notes

Keep the server isolated to trusted environments. Since VibeStudio runs as local shell commands, ensure access to the MCP client is controlled and that input paths are validated to avoid unintended file access or modifications. Regularly review the script permissions and monitor logs for unusual activity.

Sanity checks and quick verification

Run a basic initialization to confirm the server responds and exposes its tool list, then perform a small, non-destructive operation such as inspecting media metadata or generating a thumbnail from a test video. Ensure the output files are created and the results contain the expected metadata or artifacts.

Notes

No real-time streaming is supported and processing is sequential. File paths must be accessible from the server’s working directory, and large files may take longer to process. You can expand workflows by chaining multiple tools in sequence within your MCP client.

Appendix: Available commands and inputs

The following tools are exposed by the server: get_media_info, convert_video, extract_audio, trim_media, merge_videos, add_subtitles, resize_video, extract_frames, create_video_from_images, apply_video_filter, generate_thumbnail, convert_audio, normalize_audio. Each tool accepts a JSON-like argument structure with input_file, output_file, and tool-specific parameters.

Examples of common configurations

To use VibeStudio with a local MCP client, configure the stdio server entry as shown in the configuration example. Start the server process and point the MCP client to the local stdio interface. Then call tools such as get_media_info, convert_video, or trim_media with the appropriate arguments.

Available tools

get_media_info

Fetches detailed information about a media file including duration, codecs, and resolution.

convert_video

Converts a video to a different format or codec using FFmpeg.

extract_audio

Extracts the audio track from a video into a separate audio file.

trim_media

Trims video or audio to a specified start time and duration.

merge_videos

Concatenates multiple videos into a single output file.

add_subtitles

Burns subtitles directly into the video.

resize_video

Changes the video dimensions to the target width and height.

extract_frames

Exports frames from a video as image files.

create_video_from_images

Creates a video from a sequence of images.

apply_video_filter

Applies FFmpeg filters to the input video.

generate_thumbnail

Generates a thumbnail image from a video at a specific time.

convert_audio

Converts audio to a different format.

normalize_audio

Normalizes the audio volume for consistent playback.