# project-monai/monai architecture

> Architecture diagram of the GitHub repository project-monai/monai, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-10). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/project-monai/monai
- Repository: https://github.com/project-monai/monai

## Overview

MONAI is a PyTorch-based framework for medical-imaging AI. A typical user workflow loads imaging data, composes preprocessing transforms, trains or evaluates a network, and applies inference and metrics; MONAI also provides reusable losses, visualization, and specialized application workflows. The graph emphasizes those product capabilities and the metadata-aware data layer. Connections are limited to relationships documented in the README or visible in the supplied source index; unsampled application links are omitted.

Main entry points: [`monai/transforms/compose.py`](https://github.com/project-monai/monai/blob/dev/monai/transforms/compose.py) (Transform composition), [`monai/engines/trainer.py`](https://github.com/project-monai/monai/blob/dev/monai/engines/trainer.py) (Training engine), [`monai/data/dataset.py`](https://github.com/project-monai/monai/blob/dev/monai/data/dataset.py) (Datasets and readers), [`monai/data/meta_tensor.py`](https://github.com/project-monai/monai/blob/dev/monai/data/meta_tensor.py) (Metadata tensors)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Imaging data"]
  node_data_api["Datasets and readers<br/>[dataset.py]"]
  node_meta["Metadata tensors<br/>[meta_tensor.py]"]
  node_box_utils["Box utilities<br/>[box_utils.py]"]
  node_config_types["Shared type definitions"]
  node_runtime_utils["Runtime utilities<br/>[misc.py]"]
end

subgraph group_transform["Transforms"]
  node_transforms["Transform composition<br/>[compose.py]"]
  node_spatial["Spatial transforms<br/>[array.py]"]
  node_intensity["Intensity transforms<br/>[array.py]"]
  node_transform_core["Transform interfaces<br/>[transform.py]"]
end

subgraph group_model["Model development"]
  node_networks["Neural networks"]
  node_network_blocks["Network components"]
  node_losses["Training losses<br/>[dice.py]"]
end

subgraph group_workflow["Training and inference"]
  node_engine["Training engine<br/>[trainer.py]"]
  node_inferers["Inference strategies<br/>[inferer.py]"]
  node_metrics["Evaluation metrics"]
  node_visualization["Visualization"]
end

subgraph group_applications["Specialized applications"]
  node_bundle["Portable bundles<br/>[config_parser.py]"]
  node_auto3dseg["Auto3DSeg<br/>[analyzer.py]"]
  node_federated["Federated learning<br/>[client_algo.py]"]
  node_pathology["Pathology tools"]
  node_reconstruction["MRI reconstruction"]
end

node_user(("Researcher"))

node_user -->|"loads data"| node_data_api
node_data_api -->|"uses metadata"| node_meta
node_data_api -->|"uses transforms"| node_transforms
node_user -->|"preprocesses images"| node_transforms
node_transforms -->|"composes"| node_spatial
node_transforms -->|"composes"| node_intensity
node_spatial -->|"implements"| node_transform_core
node_intensity -->|"implements"| node_transform_core
node_spatial -->|"updates metadata"| node_meta
node_user -->|"builds models"| node_networks
node_networks -.->|"uses components"| node_network_blocks
node_user -->|"trains models"| node_engine
node_engine -->|"uses data"| node_data_api
node_engine -->|"uses transforms"| node_transforms
node_engine -->|"uses inferer"| node_inferers
node_user -->|"chooses objectives"| node_losses
node_user -->|"evaluates results"| node_metrics
node_user -->|"runs inference"| node_inferers
node_inferers -->|"uses networks"| node_networks
node_user -.->|"inspects outputs"| node_visualization
node_user -->|"configures workflows"| node_bundle
node_user -.->|"runs Auto3DSeg"| node_auto3dseg
node_user -->|"runs distributed learning"| node_federated
node_user -.->|"uses pathology tools"| node_pathology
node_user -.->|"uses reconstruction"| node_reconstruction
node_meta -->|"uses types"| node_config_types
node_meta -->|"uses utilities"| node_runtime_utils

click node_data_api "https://github.com/project-monai/monai/blob/dev/monai/data/dataset.py"
click node_meta "https://github.com/project-monai/monai/blob/dev/monai/data/meta_tensor.py"
click node_box_utils "https://github.com/project-monai/monai/blob/dev/monai/data/box_utils.py"
click node_transforms "https://github.com/project-monai/monai/blob/dev/monai/transforms/compose.py"
click node_spatial "https://github.com/project-monai/monai/blob/dev/monai/transforms/spatial/array.py"
click node_intensity "https://github.com/project-monai/monai/blob/dev/monai/transforms/intensity/array.py"
click node_transform_core "https://github.com/project-monai/monai/blob/dev/monai/transforms/transform.py"
click node_networks "https://github.com/project-monai/monai/tree/dev/monai/networks/nets"
click node_network_blocks "https://github.com/project-monai/monai/tree/dev/monai/networks/blocks"
click node_losses "https://github.com/project-monai/monai/blob/dev/monai/losses/dice.py"
click node_engine "https://github.com/project-monai/monai/blob/dev/monai/engines/trainer.py"
click node_inferers "https://github.com/project-monai/monai/blob/dev/monai/inferers/inferer.py"
click node_metrics "https://github.com/project-monai/monai/tree/dev/monai/metrics"
click node_visualization "https://github.com/project-monai/monai/tree/dev/monai/visualize"
click node_bundle "https://github.com/project-monai/monai/blob/dev/monai/bundle/config_parser.py"
click node_auto3dseg "https://github.com/project-monai/monai/blob/dev/monai/auto3dseg/analyzer.py"
click node_federated "https://github.com/project-monai/monai/blob/dev/monai/fl/client/client_algo.py"
click node_pathology "https://github.com/project-monai/monai/tree/dev/monai/apps/pathology"
click node_reconstruction "https://github.com/project-monai/monai/tree/dev/monai/apps/reconstruction"
click node_config_types "https://github.com/project-monai/monai/blob/dev/monai/config/type_definitions.py"
click node_runtime_utils "https://github.com/project-monai/monai/blob/dev/monai/utils/misc.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_data_api,node_meta,node_box_utils,node_config_types,node_runtime_utils toneBlue
class node_transforms,node_spatial,node_intensity,node_transform_core toneAmber
class node_networks,node_network_blocks,node_losses toneMint
class node_engine,node_inferers,node_metrics,node_visualization toneRose
class node_bundle,node_auto3dseg,node_federated,node_pathology,node_reconstruction,node_user toneIndigo
```

## Components

### Imaging data

- [Datasets and readers](https://github.com/project-monai/monai/blob/dev/monai/data/dataset.py): `monai/data/dataset.py`
- [Metadata tensors](https://github.com/project-monai/monai/blob/dev/monai/data/meta_tensor.py): `monai/data/meta_tensor.py`
- [Box utilities](https://github.com/project-monai/monai/blob/dev/monai/data/box_utils.py): `monai/data/box_utils.py`
- [Shared type definitions](https://github.com/project-monai/monai/blob/dev/monai/config/type_definitions.py): `monai/config/type_definitions.py`
- [Runtime utilities](https://github.com/project-monai/monai/blob/dev/monai/utils/misc.py): `monai/utils/misc.py`

### Transforms

- [Transform composition](https://github.com/project-monai/monai/blob/dev/monai/transforms/compose.py): `monai/transforms/compose.py`
- [Spatial transforms](https://github.com/project-monai/monai/blob/dev/monai/transforms/spatial/array.py): `monai/transforms/spatial/array.py`
- [Intensity transforms](https://github.com/project-monai/monai/blob/dev/monai/transforms/intensity/array.py): `monai/transforms/intensity/array.py`
- [Transform interfaces](https://github.com/project-monai/monai/blob/dev/monai/transforms/transform.py): `monai/transforms/transform.py`

### Model development

- [Neural networks](https://github.com/project-monai/monai/tree/dev/monai/networks/nets): `monai/networks/nets`
- [Network components](https://github.com/project-monai/monai/tree/dev/monai/networks/blocks): `monai/networks/blocks`
- [Training losses](https://github.com/project-monai/monai/blob/dev/monai/losses/dice.py): `monai/losses/dice.py`

### Training and inference

- [Training engine](https://github.com/project-monai/monai/blob/dev/monai/engines/trainer.py): `monai/engines/trainer.py`
- [Inference strategies](https://github.com/project-monai/monai/blob/dev/monai/inferers/inferer.py): `monai/inferers/inferer.py`
- [Evaluation metrics](https://github.com/project-monai/monai/tree/dev/monai/metrics): `monai/metrics`
- [Visualization](https://github.com/project-monai/monai/tree/dev/monai/visualize): `monai/visualize`

### Specialized applications

- [Portable bundles](https://github.com/project-monai/monai/blob/dev/monai/bundle/config_parser.py): `monai/bundle/config_parser.py`
- [Auto3DSeg](https://github.com/project-monai/monai/blob/dev/monai/auto3dseg/analyzer.py): `monai/auto3dseg/analyzer.py`
- [Federated learning](https://github.com/project-monai/monai/blob/dev/monai/fl/client/client_algo.py): `monai/fl/client/client_algo.py`
- [Pathology tools](https://github.com/project-monai/monai/tree/dev/monai/apps/pathology): `monai/apps/pathology`
- [MRI reconstruction](https://github.com/project-monai/monai/tree/dev/monai/apps/reconstruction): `monai/apps/reconstruction`

### Other components

- **Researcher**

## Connections

- Researcher → Datasets and readers: loads data
- Datasets and readers → Metadata tensors: uses metadata (evidence: `monai/data/dataset.py`)
- Datasets and readers → Transform composition: uses transforms (evidence: `monai/data/dataset.py`)
- Researcher → Transform composition: preprocesses images
- Transform composition → Spatial transforms: composes (evidence: `monai/transforms/__init__.py`)
- Transform composition → Intensity transforms: composes (evidence: `monai/transforms/__init__.py`)
- Spatial transforms → Transform interfaces: implements (evidence: `monai/transforms/spatial/array.py`)
- Intensity transforms → Transform interfaces: implements (evidence: `monai/transforms/intensity/array.py`)
- Spatial transforms → Metadata tensors: updates metadata (evidence: `monai/transforms/spatial/array.py`)
- Researcher → Neural networks: builds models
- Neural networks → Network components: uses components
- Researcher → Training engine: trains models
- Training engine → Datasets and readers: uses data (evidence: `monai/engines/trainer.py`)
- Training engine → Transform composition: uses transforms (evidence: `monai/engines/trainer.py`)
- Training engine → Inference strategies: uses inferer (evidence: `monai/engines/trainer.py`)
- Researcher → Training losses: chooses objectives
- Researcher → Evaluation metrics: evaluates results
- Researcher → Inference strategies: runs inference
- Inference strategies → Neural networks: uses networks (evidence: `monai/inferers/inferer.py`)
- Researcher → Visualization: inspects outputs
- Researcher → Portable bundles: configures workflows
- Researcher → Auto3DSeg: runs Auto3DSeg
- Researcher → Federated learning: runs distributed learning
- Researcher → Pathology tools: uses pathology tools
- Researcher → MRI reconstruction: uses reconstruction
- Metadata tensors → Shared type definitions: uses types (evidence: `monai/data/meta_tensor.py`)
- Metadata tensors → Runtime utilities: uses utilities (evidence: `monai/data/meta_tensor.py`)
