---

name: plugin-system description: "Guides Meshroom plugin, node-folder, template, submitter, config.json, user-plugin, Rez-plugin, and process-environment discovery workflows." disable-model-invocation: true metadata: disco-role: operating license: NOASSERTION

Meshroom Plugin System

Use this route when Meshroom cannot see a custom node/template/plugin, when plugin configuration or process environments are involved, or when a user asks how to package a Meshroom plugin.

Read First

• Plugin loading reference
• Process environments
• Troubleshooting
• Run scripts/check_plugin_folder.py before debugging a loaded provider.

Minimum Plugin Shape

plugin-root/
  meshroom/
    customNodes/
      __init__.py
      MyNode.py
    customTemplate.mg
    config.json                 # optional
  venv/                         # optional plugin runtime

Set MESHROOM_PLUGINS_PATH to plugin-root, not to the nested meshroom/ folder. Meshroom scans packages below that folder, loads node descriptor providers, registers .mg templates, and constructs a process environment.

Initialization Order

For programmatic inspection, initialize in this order when needed:

import meshroom
meshroom.setupEnvironment()
import meshroom.core
meshroom.core.initPlugins()
meshroom.core.initNodes()
meshroom.core.initPipelines()
meshroom.core.initSubmitters()

• initPlugins() loads classic/user/Rez plugin roots.
• initNodes() loads built-in nodes plus MESHROOM_NODES_PATH folders.
• initPipelines() registers templates from explicit template paths and loaded plugins.
• initSubmitters() loads submitter packages from MESHROOM_SUBMITTERS_PATH.

The CLI may choose a different order for its workflow; do not assume a template is visible before its plugin has been loaded.

Configuration Rules

config.json is an array of entries with key, type, and value. type: "path" resolves relative paths against the plugin root when the path exists. User environment variables take precedence over config defaults in the merged process environment.

Route Elsewhere

• Author the node descriptor itself with node-descriptors.
• Use a discovered template or run it from the CLI with cli-pipeline-execution.
• Debug LocalFarm daemon/task behavior with local-farm-submission.