video-editor-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| WORKSPACE_DIR | No | Relative workspace; Compose intentionally fixes this to its mounted directory | workspace |
| FFMPEG_THREADS | No | Decoder, filter and video encoder thread budget | 2 |
| MAX_FILE_BYTES | No | Maximum individual input/output file size | 536870912 |
| EXEC_TIMEOUT_MS | No | Whole-call deadline shared across copies and subprocesses | 180000 |
| MAX_RENDER_SECONDS | No | Maximum whole-video edit/QA and composition duration | 300 |
| MAX_COMPOSITION_SHOTS | No | Maximum shot/concat count | 32 |
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 | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| probeC | Inspect duration, tracks, codecs, resolution and frame rate. |
| trimB | Trim video. Auto attempts stream copy, then falls back to encoding; copy is keyframe-aligned. |
| concatA | Concatenate complete clips with normalized geometry, frame rate and audio; re-encodes. |
| transformB | Resize, crop or pad a video to a ratio or explicit even dimensions. |
| overlayB | Burn typed text or a local raster image into a video during a time window. |
| audio_mixC | Mix music under source audio or a separate voice track, with gain, fades and optional loudness normalization. |
| renderB | Render ordered shots, local overlays, per-shot audio and optional music to MP4. |
| qaC | Measure duration, black intervals, silence, integrated LUFS, resolution and FPS. Failed analysis is an error, not a pass. |
| healthA | Report ffmpeg/ffprobe versions, available encoders and workspace free space. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 9 tools
Each tool targets a distinct editing operation, and most are easy to tell apart. The only real ambiguity is probe vs qa, since both report duration/resolution/FPS, but qa is clearly a quality gate while probe focuses on codecs and tracks.
Most names are readable lowercase verbs like trim, concat, transform, and render, but the set mixes bare verbs, the underscored compound audio_mix, and noun-style names qa and health. There is no consistent verb_noun or noun_verb convention across the full set.
Nine tools is well within the ideal 3-15 range for a focused server. Each tool earns its place by covering a distinct stage of a video editing pipeline without redundancy or bloat.
The surface covers the full lifecycle: inspect, trim, assemble, transform, overlay, mix audio, render, QA, and environment health. There are no obvious dead ends for a video-editing workflow; render and qa close the loop.