# microsoft/responsible-ai-toolbox architecture

> Architecture diagram of the GitHub repository microsoft/responsible-ai-toolbox, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-09-25). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/microsoft/responsible-ai-toolbox
- Repository: https://github.com/microsoft/responsible-ai-toolbox

## Overview

Responsible AI Toolbox provides interactive interfaces and libraries for assessing model errors and fairness, interpreting predictions, and supporting data-driven decisions. The principal workflow is to open an assessment experience, explore model or cohort behavior, and investigate individual predictions or possible interventions. Shared UI components support cohort and chart interactions; Python managers and the error-analysis package provide analysis capabilities. README capabilities are included where implementation was not sampled, so some integration wiring is intentionally left unspecified.

Main entry points: [`apps/dashboard/src/app/App.tsx`](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/app/App.tsx) (RAI dashboard), [`apps/dashboard/src/model-assessment/App.tsx`](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment/App.tsx) (Model assessment), [`erroranalysis/erroranalysis/analyzer/error_analyzer.py`](https://github.com/microsoft/responsible-ai-toolbox/blob/main/erroranalysis/erroranalysis/analyzer/error_analyzer.py) (Error analysis engine), [`apps/dashboard/src/model-assessment-vision/App.tsx`](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment-vision/App.tsx) (Vision assessment)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_experience["User experience"]
  node_dashboard["RAI dashboard<br/>[App.tsx]"]
  node_registry["App datasets<br/>[applications.ts]"]
  node_widget["Embeddable widget<br/>[App.tsx]"]
end

subgraph group_assessment["Model assessment"]
  node_model_assessment["Model assessment<br/>[App.tsx]"]
  node_forecasting["Forecast assessment<br/>[App.tsx]"]
  node_error_ui["Error analysis UI<br/>[App.tsx]"]
  node_fairness_ui["Fairness UI<br/>[App.tsx]"]
  node_text_assessment["Text assessment<br/>[App.tsx]"]
  node_vision_assessment["Vision assessment<br/>[App.tsx]"]
end

subgraph group_explanation["Interpretability"]
  node_interpret_ui["Interpretability UI<br/>[App.tsx]"]
  node_text_interpret["Text explanations<br/>[App.tsx]"]
end

subgraph group_decisions["Decision support"]
  node_counterfactual["Counterfactual manager"]
  node_causal["Causal manager<br/>[causal_manager.py]"]
  node_causal_ui["Causal analysis UI"]
end

subgraph group_runtime["Analysis runtime"]
  node_core_ui["Shared UI components<br/>[Cohort.ts]"]
  node_error_engine["Error analysis engine<br/>[error_analyzer.py]"]
  node_text_error["Text error manager"]
  node_vision_error["Vision error manager"]
  node_text_explainer["Text explainer"]
  node_vision_explainer["Vision explainer"]
  node_error_package["Error analysis library<br/>[__init__.py]"]
end

node_user(("Practitioner"))

node_user -->|"opens"| node_dashboard
node_user -->|"embeds"| node_widget
node_dashboard -->|"selects datasets"| node_registry
node_dashboard -->|"presents assessment"| node_model_assessment
node_dashboard -->|"presents errors"| node_error_ui
node_dashboard -->|"presents fairness"| node_fairness_ui
node_dashboard -->|"presents explanations"| node_interpret_ui
node_dashboard -->|"presents text"| node_text_interpret
node_model_assessment -->|"uses shared UI"| node_core_ui
node_forecasting -->|"uses shared UI"| node_core_ui
node_error_ui -->|"requests analysis"| node_error_package
node_text_assessment -.->|"supports text analysis"| node_text_error
node_vision_assessment -.->|"supports vision analysis"| node_vision_error
node_text_interpret -.->|"obtains explanations"| node_text_explainer
node_vision_assessment -.->|"obtains explanations"| node_vision_explainer
node_model_assessment -.->|"requests counterfactuals"| node_counterfactual
node_model_assessment -.->|"presents causal insights"| node_causal_ui
node_causal_ui -.->|"requests causal effects"| node_causal
node_error_package -->|"provides analysis"| node_error_engine
node_error_engine -.->|"supports text analysis"| node_text_error
node_error_engine -.->|"supports vision analysis"| node_vision_error

click node_dashboard "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/app/App.tsx"
click node_registry "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/app/applications.ts"
click node_widget "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/widget/src/app/App.tsx"
click node_model_assessment "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment/App.tsx"
click node_forecasting "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment-forecasting/App.tsx"
click node_error_ui "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/error-analysis/App.tsx"
click node_fairness_ui "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/fairness/App.tsx"
click node_text_assessment "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment-text/App.tsx"
click node_vision_assessment "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment-vision/App.tsx"
click node_interpret_ui "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/interpret/App.tsx"
click node_text_interpret "https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/interpret-text/App.tsx"
click node_core_ui "https://github.com/microsoft/responsible-ai-toolbox/blob/main/libs/core-ui/src/lib/Cohort/Cohort.ts"
click node_error_engine "https://github.com/microsoft/responsible-ai-toolbox/blob/main/erroranalysis/erroranalysis/analyzer/error_analyzer.py"
click node_text_error "https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_text/responsibleai_text/managers/error_analysis_manager.py"
click node_vision_error "https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_vision/responsibleai_vision/managers/error_analysis_manager.py"
click node_text_explainer "https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_text/responsibleai_text/managers/explainer_manager.py"
click node_vision_explainer "https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_vision/responsibleai_vision/managers/explainer_manager.py"
click node_counterfactual "https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai/responsibleai/managers/counterfactual_manager.py"
click node_causal "https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai/responsibleai/managers/causal_manager.py"
click node_causal_ui "https://github.com/microsoft/responsible-ai-toolbox/blob/main/libs/causality/src/lib/CausalAnalysisDashboard/CausalInsightsTab.tsx"
click node_error_package "https://github.com/microsoft/responsible-ai-toolbox/blob/main/erroranalysis/erroranalysis/__init__.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_dashboard,node_registry,node_widget toneBlue
class node_model_assessment,node_forecasting,node_error_ui,node_fairness_ui,node_text_assessment,node_vision_assessment toneAmber
class node_interpret_ui,node_text_interpret toneMint
class node_counterfactual,node_causal,node_causal_ui toneRose
class node_core_ui,node_error_engine,node_text_error,node_vision_error,node_text_explainer,node_vision_explainer,node_error_package,node_user toneIndigo
```

## Components

### User experience

- [RAI dashboard](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/app/App.tsx): `apps/dashboard/src/app/App.tsx`
- [App datasets](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/app/applications.ts): `apps/dashboard/src/app/applications.ts`
- [Embeddable widget](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/widget/src/app/App.tsx): `apps/widget/src/app/App.tsx`

### Model assessment

- [Model assessment](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment/App.tsx): `apps/dashboard/src/model-assessment/App.tsx`
- [Forecast assessment](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment-forecasting/App.tsx): `apps/dashboard/src/model-assessment-forecasting/App.tsx`
- [Error analysis UI](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/error-analysis/App.tsx): `apps/dashboard/src/error-analysis/App.tsx`
- [Fairness UI](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/fairness/App.tsx): `apps/dashboard/src/fairness/App.tsx`
- [Text assessment](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment-text/App.tsx): `apps/dashboard/src/model-assessment-text/App.tsx`
- [Vision assessment](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/model-assessment-vision/App.tsx): `apps/dashboard/src/model-assessment-vision/App.tsx`

### Interpretability

- [Interpretability UI](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/interpret/App.tsx): `apps/dashboard/src/interpret/App.tsx`
- [Text explanations](https://github.com/microsoft/responsible-ai-toolbox/blob/main/apps/dashboard/src/interpret-text/App.tsx): `apps/dashboard/src/interpret-text/App.tsx`

### Decision support

- [Counterfactual manager](https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai/responsibleai/managers/counterfactual_manager.py): `responsibleai/responsibleai/managers/counterfactual_manager.py`
- [Causal manager](https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai/responsibleai/managers/causal_manager.py): `responsibleai/responsibleai/managers/causal_manager.py`
- [Causal analysis UI](https://github.com/microsoft/responsible-ai-toolbox/blob/main/libs/causality/src/lib/CausalAnalysisDashboard/CausalInsightsTab.tsx): `libs/causality/src/lib/CausalAnalysisDashboard/CausalInsightsTab.tsx`

### Analysis runtime

- [Shared UI components](https://github.com/microsoft/responsible-ai-toolbox/blob/main/libs/core-ui/src/lib/Cohort/Cohort.ts): `libs/core-ui/src/lib/Cohort/Cohort.ts`
- [Error analysis engine](https://github.com/microsoft/responsible-ai-toolbox/blob/main/erroranalysis/erroranalysis/analyzer/error_analyzer.py): `erroranalysis/erroranalysis/analyzer/error_analyzer.py`
- [Text error manager](https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_text/responsibleai_text/managers/error_analysis_manager.py): `responsibleai_text/responsibleai_text/managers/error_analysis_manager.py`
- [Vision error manager](https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_vision/responsibleai_vision/managers/error_analysis_manager.py): `responsibleai_vision/responsibleai_vision/managers/error_analysis_manager.py`
- [Text explainer](https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_text/responsibleai_text/managers/explainer_manager.py): `responsibleai_text/responsibleai_text/managers/explainer_manager.py`
- [Vision explainer](https://github.com/microsoft/responsible-ai-toolbox/blob/main/responsibleai_vision/responsibleai_vision/managers/explainer_manager.py): `responsibleai_vision/responsibleai_vision/managers/explainer_manager.py`
- [Error analysis library](https://github.com/microsoft/responsible-ai-toolbox/blob/main/erroranalysis/erroranalysis/__init__.py): `erroranalysis/erroranalysis/__init__.py`

### Other components

- **Practitioner**

## Connections

- Practitioner → RAI dashboard: opens
- Practitioner → Embeddable widget: embeds
- RAI dashboard → App datasets: selects datasets
- RAI dashboard → Model assessment: presents assessment
- RAI dashboard → Error analysis UI: presents errors
- RAI dashboard → Fairness UI: presents fairness
- RAI dashboard → Interpretability UI: presents explanations
- RAI dashboard → Text explanations: presents text
- Model assessment → Shared UI components: uses shared UI
- Forecast assessment → Shared UI components: uses shared UI
- Error analysis UI → Error analysis library: requests analysis
- Text assessment → Text error manager: supports text analysis
- Vision assessment → Vision error manager: supports vision analysis
- Text explanations → Text explainer: obtains explanations
- Vision assessment → Vision explainer: obtains explanations
- Model assessment → Counterfactual manager: requests counterfactuals
- Model assessment → Causal analysis UI: presents causal insights
- Causal analysis UI → Causal manager: requests causal effects
- Error analysis library → Error analysis engine: provides analysis
- Error analysis engine → Text error manager: supports text analysis
- Error analysis engine → Vision error manager: supports vision analysis