mcp-video-studio
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| remove_silenceA | Cut silent gaps from a video or audio file while keeping video and audio perfectly in sync. Args: input_path: Full file path to the source video or audio. output_path: Optional explicit output path. Defaults to '{name}_no_silence.{ext}'. db_threshold: Volume threshold in dB below which sound is considered silence (e.g. -30.0). min_silence_duration: Minimum silence length in seconds to trigger a cut (default: 0.5). padding: Seconds of voice buffer retained before and after speech to avoid clipping (default: 0.1). Returns: JSON string containing cut statistics, durations, reduction percentage, and output path. |
| smart_crop_verticalA | Transform horizontal 16:9 video into 9:16 vertical Shorts, Reels, or TikTok format. Args: input_path: Full file path to the source video. output_path: Optional output path. Defaults to '{name}vertical{mode}.{ext}'. mode: Framing method: - 'blur_pad': (Default) Blurred video background with centered sharp original. - 'center_crop': Direct 9:16 crop from the center. - 'fit_black_bars': Scale to fit with black letterbox padding. target_width: Canvas width (default: 1080). target_height: Canvas height (default: 1920). Returns: JSON string with framing details, target resolution, and output file path. |
| apply_lut_gradeA | Apply professional 3D Look-Up Table (.cube or .3dl) color grading to a video. Args: input_path: Full file path to the input video. lut_path: Path to the .cube or .3dl LUT file. output_path: Optional output path. Defaults to '{name}_graded.{ext}'. intensity: Grading strength multiplier from 0.0 (original) to 1.0 (100% LUT grade). Returns: JSON string with LUT details, intensity applied, and output file path. |
| extract_audio_and_metadataA | Extract complete media metadata (FPS, resolution, bitrate, codecs) and optionally extract audio. Args: input_path: Path to the media file. extract_audio: Whether to rip the audio track to a standalone file (default: True). audio_format: Format for extracted audio ('wav', 'mp3', 'aac', 'flac'). output_audio_path: Optional destination path for extracted audio. Returns: JSON string containing comprehensive audio/video stream parameters and audio file path. |
| generate_capcut_draftA | Generate native CapCut Desktop draft_content.json and draft_meta_info.json project files. Directly importable into CapCut Desktop / Jianying Pro without manual timeline setup. Args: project_name: Name of the project (e.g. 'TikTok_Highlights_01'). clips: List of video clip definitions with 'path', optional 'start_sec', 'end_sec', 'speed', 'volume'. audio_tracks: Optional list of background music or sound effects with 'path', 'timeline_start_sec', 'volume'. output_dir: Base directory where project will be saved. Defaults to CapCut's native drafts directory. width: Canvas width (default: 1080). height: Canvas height (default: 1920). fps: Timeline frame rate (default: 30.0). ratio: Aspect ratio ('9:16', '16:9', '1:1', 'original'). Returns: JSON string with project directory, JSON file paths, timeline duration, and opening instructions. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| get_system_info | Retrieve host environment status, FFmpeg binaries, and CapCut drafts location. |
| get_presets | Retrieve standard video formatting and silence removal presets. |
TDQS
Scored across 5 tools
Each tool targets a clearly distinct operation: silence removal, vertical reframing, LUT color grading, metadata/audio extraction, and CapCut draft generation. There is essentially no overlap in purpose or input semantics, so an agent can select the right tool unambiguously.
All names use consistent snake_case with verb-led phrasing (remove_silence, apply_lut_grade, extract_audio_and_metadata, generate_capcut_draft), with smart_crop_vertical being the only minor adjective-led variant. The convention is predictable and readable throughout.
Five tools is a reasonable, well-scoped set where each earns its place in a short-form video pipeline. It leans slightly thin for a 'video studio' that could plausibly offer basic trim/concat, but nothing is bloated or redundant.
The tools cover a coherent shorts-production workflow (inspect → silence-cut → vertical crop → grade → assemble CapCut project). Minor gaps exist around primitive editing operations like trimming, concatenation, or subtitle generation, but these can be worked around via the CapCut draft output.