MCP 클라이언트가 도구와 리소스를 통해 DaVinci Resolve(Fusion 포함)를 제어하고 검사할 수 있게 해주는 Model Context Protocol(MCP) 서버입니다.
프로젝트 관리 (생성, 로드, 저장)
타임라인 작업 (타임라인 생성, 타임라인 전환, 클립으로 구성)
미디어 풀 작업 (미디어 가져오기, 폴더 생성, 폴더 검사)
Fusion 도우미 (클립에 컴포지션 추가, 노드 생성, 노드 체인 생성)
Resolve 페이지 탐색
고급 스크립팅 (execute_python, execute_lua)
Related MCP server: Resolve Claude MCP
Python 3.10 이상
스크립팅 API가 활성화된 DaVinci Resolve
MCP 호환 클라이언트 (예: Claude Desktop)
git clone https://github.com/apvlv/davinci-resolve-mcp.git
cd davinci-resolve-mcp
uv sync --all-extras --devuv run davinci-resolve-mcp서버는 시작 시 Resolve에 연결을 시도합니다. DaVinci Resolve가 실행 중인지 확인하십시오.
MCP 클라이언트가 서버 명령을 실행하도록 구성하십시오:
{
"mcpServers": {
"davinci-resolve": {
"command": "uv",
"args": ["run", "davinci-resolve-mcp"]
}
}
}system://status
project://current
project://timelines
timeline://current
mediapool://folders
mediapool://current
storage://volumes
create_project(name)
load_project(name)
save_project()
create_timeline(name)
set_current_timeline(index)
import_media(file_paths)
create_folder(name)
create_timeline_from_clips(name, clip_indices)
add_fusion_comp_to_clip(timeline_index, track_type, track_index, item_index)
create_fusion_node(node_type, parameters)
create_fusion_node_chain(node_chain)
open_page(page_name) (페이지 이름은 다음 중 하나여야 합니다:
media, edit, fusion, color, fairlight, deliver)
execute_python(code)
execute_lua(script)
make install # install deps and pre-commit hooks
make lint-check # ruff check
make format-check # ruff format check
make test # pytest with coverage
make check # lint + format + tests"My Documentary라는 이름의 프로젝트를 생성해줘"
"현재 프로젝트의 타임라인 목록을 보여줘"
"이 파일들을 미디어 풀로 가져와줘"
"Color 페이지를 열어줘"
"Blur -> ColorCorrector로 이어지는 Fusion 노드 체인을 만들어줘"
MIT (LICENSE 참조)
Add a Fusion composition to a clip in the timeline.
Args: timeline_index: The index of the timeline (1-based) track_type: The type of track ("video", "audio", or "subtitle") track_index: The index of the track (1-based) item_index: The index of the item in the track (1-based)
Returns: A message indicating success or failure
Inputs: timeline_index (required), track_type (required), track_index (required), item_index (required). Returns: result
Create a new folder in the current media pool folder.
Args: name: The name of the folder to create
Returns: A message indicating success or failure
Inputs: name (required). Returns: result
Create a Fusion node in the current composition.
Args: node_type: The type of node to create (e.g., 'Blur', 'ColorCorrector', 'Text') parameters: Optional dictionary of parameters to set on the node
Returns: A message indicating success or failure
Inputs: node_type (required), parameters. Returns: result
Create a chain of connected Fusion nodes in the current composition.
Args: node_chain: A list of dictionaries, each containing: - 'type': The type of node to create - 'name': Optional name for the node - 'params': Optional dictionary of parameters to set on the node
Returns: A message indicating success or failure
Inputs: node_chain (required). Returns: result
Create a new DaVinci Resolve project.
Args: name: The name of the project to create
Returns: A message indicating success or failure
Inputs: name (required). Returns: result
Create a new timeline in the current project.
Args: name: The name of the timeline to create
Returns: A message indicating success or failure
Inputs: name (required). Returns: result
Create a new timeline from clips in the current media pool folder.
Args: name: The name of the timeline to create clip_indices: A list of clip indices (1-based) to include in the timeline
Returns: A message indicating success or failure
Inputs: name (required), clip_indices (required). Returns: result
Execute a Lua script in DaVinci Resolve's Fusion.
Args: script: The Lua script to execute
Returns: The result of the script execution
Inputs: script (required). Returns: result
Execute arbitrary Python code in DaVinci Resolve.
Args: code: The Python code to execute
Returns: The result of the code execution
Inputs: code (required). Returns: result
Import media files into the current media pool folder.
Args: file_paths: A list of file paths to import
Returns: A message indicating success or failure
Inputs: file_paths (required). Returns: result
Load an existing DaVinci Resolve project.
Args: name: The name of the project to load
Returns: A message indicating success or failure
Inputs: name (required). Returns: result
Open a specific page in DaVinci Resolve.
Args: page_name: The name of the page to open (media, edit, fusion, color, fairlight, deliver)
Returns: A message indicating success or failure
Inputs: page_name (required). Returns: result
Save the current DaVinci Resolve project.
Returns: A message indicating success or failure
Inputs: no named parameters. Returns: result
Set the current timeline by index.
Args: index: The index of the timeline (1-based)
Returns: A message indicating success or failure
Inputs: index (required). Returns: result
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 14 tools
Each tool has a clearly distinct purpose targeting specific DaVinci Resolve operations like timeline management, Fusion composition, media import, and scripting. No tools overlap in functionality, making selection straightforward for an agent.
All tool names follow a consistent verb_noun pattern using snake_case, such as 'create_project', 'import_media', and 'execute_lua'. This uniformity enhances readability and predictability across the toolset.
With 14 tools, the server is well-scoped for DaVinci Resolve operations, covering project management, timeline creation, media handling, Fusion nodes, and scripting. Each tool serves a clear purpose without redundancy or bloat.
The toolset provides comprehensive coverage for core workflows including project lifecycle, timeline operations, media import, and Fusion composition. Minor gaps exist, such as the lack of tools for deleting or modifying existing projects or timelines, but agents can work around these with available tools.
Direct AI video, image, music and storyboard production from any MCP client.
Use AI models for chat, image, and video generation from Claude Code and other MCP hosts.
AI video editor for agents and humans: timeline, captions, color, audio and generation as MCP tools.
The agentic video editor for ChatGPT and Claude. Ask in your chat and it cuts, captions, adds music, titles and motion graphics on a real timeline you can open and adjust.
