Node-Based Workflow Authoring
- Visually design your data processing pipelines.
- Connect operations to define dependencies, data flow, and execution order.
- An intuitive graph interface to zoom, drag and drop files, connect nodes, and visualize computation progress.
Modular & Extensible Architecture
- Every operation is encapsulated as a node, making it easy to add, replace, or reuse components.
- Supports dynamic loading of plugins to introduce new node types without modifying the core.
- Ideal for creating domain-specific toolsets (e.g., AI, vision, simulation, data prep).
Cross-Platform & Open
- Prebuilt versions for Linux and Windows, but should be packaged manually for macOS.
- Fully open-source and scriptable, allowing tight integration with any environment or CI system.
- Consistent behavior across platforms for reproducible workflows.
Scriptable & Automatable
- Command-line interface (
meshroom_batch) for headless execution and integration into automated pipelines. - JSON-based pipeline definitions for reproducibility and version control.
- Easily combine visual editing and scripting for hybrid workflows.
Live Debugging & Monitoring
- Inspect intermediate results directly in the UI.
- Re-run individual nodes or subgraphs without restarting the full pipeline.
- Real-time feedback and error reporting make debugging fast and intuitive.
Flexible Data Handling
- Caching system reuse computed results to speed up experimentations on compute intensive pipelines.
- Connect structured data between nodes.
- Dedicated cache folder per node instance.
Reusable Templates
- Save and share graph templates for recurring workflows.
- Parameter presets and node groups simplify standardization across teams.
- Quickly clone, modify, or chain templates to build larger workflows.
- Pipelines can be serialized, versioned, and shared.
Integrated Viewers
- Built-in viewers for logs, images and 3D data.
- Inspection of node outputs.
Isolated Execution Environment
- Each plugin can run in its own virtual environment for dependency isolation.
- Facilitates reproducible research and portable workflows.
- Clear dependency tracking ensures results can be replicated later.
- Perfect for scientific, industrial, or research environments requiring auditability.
Scalable Execution
- Run workflows locally or on distributed systems.
- On distributed systems, automatically selects the best hardware based on each node’s requirements.
- Compute simultaneously on the same project, both locally and remotely.
- Plugin-based submitters allow integration with new render farm or compute solutions.
- Scales seamlessly from small scripts to large, data-intensive workflows.
Collaborative & Open Ecosystem
- Promotes community contributions, peer review, and shared development.
- Uses the MPL license to encourage sharing and contributions while allowing commercial use, fostering collaboration between industry and academia.