CapCut MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PYTHONPATH | No | Path to the capcut-mcp/src directory, used in the MCP server env configuration so Python can locate the capcut_mcp package. |
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
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| capcut_improve_audioB | Studio-grade vocal enhancement and loudness mastering for UGC videos.
|
| capcut_create_draftB | Create a new vertical (9:16) or landscape CapCut project draft on disk. Draft will be created in %LOCALAPPDATA%\CapCut\User Data\Projects\com.lveditor.draft. |
| capcut_add_video_clipC | Add a video or photo segment to the CapCut timeline.
|
| capcut_set_background_removalB | Toggle background removal on an existing video clip in the CapCut draft.
|
| capcut_add_audio_clipC | Add an audio clip (clean voiceover, background music, SFX) to the CapCut draft. |
| capcut_add_punch_zoomsB | Add dynamic punch zooms (alternating between 1.0x and zoom_scale every interval_s) to a talking-head clip to maintain high viewer retention in short-form ads. |
| capcut_add_bg_music_with_duckingB | Add background music with automatic volume ducking under speech.
|
| capcut_add_captionsC | Add styled captions to the CapCut timeline. captions_json format: [{"text": "...", "start_ms": 0, "duration_ms": 1500}] |
| capcut_add_karaoke_captionsC | Add viral Hormozi-style karaoke captions where the active spoken word lights up in bright gold/yellow (#FFD700) or electric green (#00FF66). word_timestamps_json: JSON string of [{"word": "Stop", "start_ms": 0, "end_ms": 250}, ...] |
| capcut_add_proof_overlayB | Add an Amazon review screenshot, PubMed study proof, or TikTok comment bubble overlay.
|
| capcut_add_hook_badgeC | Add a high-contrast headline banner (e.g. '3 WARNING SIGNS' or 'DOCTOR WARNS') at top center above the avatar.
|
| capcut_lint_safe_zonesA | Scan the draft to verify that captions, overlays, and badges do not collide with TikTok or Instagram Reels UI elements (right-side icons, bottom sound title). Also checks for missing media files on disk. |
| capcut_clone_and_revoiceA | Clone a finished English draft into a Spanish (or localized) version: preserves all visual cuts, B-roll timing, and cutout settings, while swapping the voiceover audio and updating the captions. |
| capcut_render_preview_proxyA | Render a fast, lightweight 540x960 proxy preview MP4 directly from the draft spec using FFmpeg (~2-4 seconds), allowing agents/users to view the edit without launching CapCut. |
| capcut_assemble_ugc_videoA | One-shot UGC assembly:
|
| capcut_list_draftsA | List all CapCut drafts found on this machine. |
| capcut_inspect_draftB | Inspect timeline tracks, segments, and materials of a draft. |
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 17 tools
Most tools target distinct operations (draft CRUD, clip/audio/caption insertion, audio mastering, background removal, linting, preview render). The only real overlap is between the one-shot capcut_assemble_ugc_video and the granular capcut_add_* / capcut_improve_audio / capcut_set_background_removal tools, but the 'one-shot' framing makes the boundary readable.
Every tool uses the capcut_ prefix followed by a snake_case verb_noun pattern (list_drafts, add_captions, set_background_removal, render_preview_proxy). No camelCase or mixed conventions; highly predictable.
17 tools is slightly above the ideal band but each maps to a genuine, non-redundant editing operation in a fairly complex video-automation domain. It leans heavy but is defensible rather than bloated.
The surface covers draft creation/inspection, media insertion, captions, overlays, audio mastering, localization, linting, and preview rendering. Notable gaps are delete operations (no remove_clip/delete_draft) and full final export, though CapCut Desktop is intended to handle export, so agents can work around these.