dcc-mcp-premiere

Typed DCC-MCP control for Adobe Premiere Pro through the shared adobepy broker and Adobe's UXP runtime. The adapter exposes bounded project, media, timeline, marker, save, frame-export, and AME queue operations. It deliberately does not expose raw JavaScript, evalJs, shell commands, or arbitrary UXP calls.

Illustrative workflow generated with OpenAI ImageGen from the retained source in docs/images/sources; it is not a Premiere Pro screenshot or host-validation artifact.

Install

python -m pip install dcc-mcp-premiere
adobepy install-bridge premiere --dest <plugin-dir> --token <non-default-token>

For development or an Internal deployment with an approved prebuilt UXP bridge, the shared CLI can link that bridge into the Adobe debug-plugin directory:

dcc-mcp-cli install --dcc-type premiere `
  --plugin-source F:\studio\artifacts\premiere-uxp-bridge `
  --adobe-debug-root F:\studio\adobe-debug `
  --execute

DCC_MCP_PLUGIN_SOURCE and DCC_MCP_ADOBE_DEBUG_ROOT can be used in a studio profile. The source directory must be the bridge root and contain its manifest.json; the command does not assume that the adapter repository itself contains UXP assets. It creates an idempotent directory link, does not copy the bridge, and does not require UXP Developer Tool. Enable Premiere's UXP debug mode once, then restart Premiere after the link is created. Premiere Pro 25.6 or later is required. Set ADOBEPY_TOKEN to the bridge token, start the adapter, then verify the connected host through DCC-MCP discovery:

dcc-mcp-cli wait-ready --dcc-type premiere --timeout-secs 60
dcc-mcp-cli load-skill premiere-project --dcc-type premiere

UXP Developer Tool remains a fallback for investigating a host-specific load problem.

Each adapter instance uses an OS-assigned port and registers with DCC-MCP discovery. Agents should connect through the stable local gateway at http://127.0.0.1:9765/mcp. Set DCC_MCP_PREMIERE_PORT only for a deliberately fixed direct endpoint.

Typed tools

Inspection:

• get_status
• inspect_project
• list_sequences
• inspect_sequence
• list_project_items
• list_selected_clips
• list_encoder_presets

Project and timeline authoring:

• create_bin
• import_media
• create_sequence
• insert_project_item
• overwrite_project_item
• create_marker

Persistence and export:

• save_project
• save_project_as
• queue_sequence_export
• export_frame

List and scan operations are paginated and bounded. Imports accept at most 100 files, with per-file and aggregate byte limits. Track indices, marker text, frame dimensions, and output extensions are validated before invoking the host. AME export reports queued=true; completion must be verified separately.

Safe local paths

The adapter resolves paths before calling Premiere and confines them to:

• DCC_MCP_PREMIERE_ALLOWED_INPUT_ROOTS for imported media
• DCC_MCP_PREMIERE_ALLOWED_OUTPUT_ROOTS for .prproj and exported media
• DCC_MCP_PREMIERE_ALLOWED_PRESET_ROOTS for .sqpreset and .epr files

Each variable uses the platform path separator and defaults to the current user's home directory. Existing outputs require overwrite=true; missing parent directories require create_parents=true. Verified synchronous outputs include byte count and SHA-256.

Real-host acceptance

Automated tests use contract-compatible facade fakes; they are not represented as live Premiere proof. With a disposable project open and the UXP bridge connected, configure the three allowlists and run:

set DCC_MCP_PREMIERE_SMOKE_MEDIA=C:\path\to\media.mp4
set DCC_MCP_PREMIERE_SMOKE_ROOT=C:\path\to\writable-evidence
python tools/live_premiere_smoke.py

The smoke script uses dcc-mcp-cli for readiness, skill loading, and every typed call. It imports media, authors and inspects a sequence, adds a marker, saves a verified project copy, exports a verified frame, and prints hashes. Set DCC_MCP_PREMIERE_SMOKE_EPR to an allowlisted .epr file to include an AME queue test.

Development

python -m pip install -e ".[dev]"
python -m pytest -q
python -m ruff check src tests tools
python -m ruff format --check src tests tools
python tools/lint_skills.py
python -m build
python -m twine check dist/*

See docs/architecture.md for ownership, readiness, and security boundaries.