# nvidia-ai-blueprints/rag architecture

> Architecture diagram of the GitHub repository nvidia-ai-blueprints/rag, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-09-27). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/nvidia-ai-blueprints/rag
- Repository: https://github.com/nvidia-ai-blueprints/rag

## Overview

NVIDIA RAG Blueprint combines a reference chat and collection-management UI with API services for document ingestion and grounded question answering. The principal workflow is document upload and multimodal extraction into a pluggable vector store, followed by query processing, retrieval/reranking, and answer generation. Agentic planning, multimodal answering, reflection, summaries, safety controls, and observability extend that flow. The README documents these capabilities; source excerpts are partial, so links between unsampled internals are limited to documented responsibilities.

Main entry points: [`src/nvidia_rag/rag_server/main.py`](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/main.py) (RAG API), [`src/nvidia_rag/ingestor_server/main.py`](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/main.py) (Ingestion API), [`frontend/src/App.tsx`](https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/App.tsx) (Reference UI), [`src/nvidia_rag/utils/vdb/vdb_base.py`](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/vdb/vdb_base.py) (Vector store)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_experience["User Experience"]
  node_ui["Reference UI<br/>[App.tsx]"]
  node_chat["Chat workflow<br/>[Chat.tsx]"]
  node_collection_ui["Collection management<br/>[NewCollection.tsx]"]
  node_settings_ui["Settings<br/>[SettingsPage.tsx]"]
end

subgraph group_ingestion["Ingestion"]
  node_ingest_api["Ingestion API<br/>[main.py]"]
  node_ingest_state["Ingestion task state"]
  node_extraction["Document extraction<br/>[nvingest.py]"]
  node_schema["Ingest schema adapter"]
  node_summarization["Document summaries<br/>[summarization.py]"]
end

subgraph group_query["Query and Retrieval"]
  node_rag_api["RAG API<br/>[main.py]"]
  node_query_processing["Query processing"]
  node_agentic["Agentic RAG<br/>[agentic_rag.py]"]
  node_retrieval["Retrieval and reranking<br/>[reranker.py]"]
end

subgraph group_answer["Answering"]
  node_generation["Response generation"]
  node_vlm["Multimodal answering<br/>[vlm.py]"]
  node_reflection["Reflection<br/>[reflection.py]"]
  node_safety["Optional guardrails"]
end

subgraph group_platform["Storage and Operations"]
  node_vector_store[("Vector store<br/>[vdb_base.py]")]
  node_elasticsearch[("Elasticsearch adapter<br/>[elastic_vdb.py]")]
  node_milvus[("Milvus adapter<br/>[milvus_vdb.py]")]
  node_nim_models{{"NVIDIA NIM services"}}
  node_telemetry["Telemetry and tracing<br/>[instrumentation.py]"]
end

node_user(("User"))
node_documents["Enterprise documents"]
node_api_client(("API client"))

node_user -->|"uses"| node_ui
node_ui -->|"routes to"| node_chat
node_ui -->|"routes to"| node_collection_ui
node_ui -->|"routes to"| node_settings_ui
node_chat -->|"submits query"| node_rag_api
node_api_client -->|"calls API"| node_rag_api
node_rag_api -->|"processes query"| node_query_processing
node_query_processing -->|"prepares search"| node_retrieval
node_rag_api -.->|"dispatches agentic"| node_agentic
node_agentic -->|"executes searches"| node_retrieval
node_retrieval -->|"searches"| node_vector_store
node_elasticsearch -->|"implements"| node_vector_store
node_milvus -->|"implements"| node_vector_store
node_rag_api -->|"generates answer"| node_generation
node_generation -->|"requests inference"| node_nim_models
node_rag_api -.->|"dispatches multimodal"| node_vlm
node_vlm -.->|"requests vision inference"| node_nim_models
node_rag_api -.->|"invokes reflection"| node_reflection
node_rag_api -.->|"filters content"| node_safety
node_documents -->|"uploads"| node_ingest_api
node_user -->|"configures collection"| node_collection_ui
node_collection_ui -->|"creates collection"| node_ingest_api
node_ingest_api -->|"tracks progress"| node_ingest_state
node_ingest_api -->|"dispatches extraction"| node_extraction
node_extraction -->|"extracts content"| node_nim_models
node_extraction -->|"adapts records"| node_schema
node_schema -->|"writes records"| node_vector_store
node_ingest_api -.->|"requests summaries"| node_summarization
node_rag_api -->|"emits traces"| node_telemetry
node_ingest_api -->|"emits traces"| node_telemetry

click node_ui "https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/App.tsx"
click node_chat "https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/pages/Chat.tsx"
click node_collection_ui "https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/pages/NewCollection.tsx"
click node_settings_ui "https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/pages/SettingsPage.tsx"
click node_rag_api "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/main.py"
click node_query_processing "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/query_decomposition.py"
click node_agentic "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/agentic_rag/agentic_rag.py"
click node_retrieval "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/reranker.py"
click node_vector_store "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/vdb/vdb_base.py"
click node_elasticsearch "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/vdb/elasticsearch/elastic_vdb.py"
click node_milvus "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/vdb/milvus/milvus_vdb.py"
click node_generation "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/response_generator.py"
click node_vlm "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/vlm.py"
click node_reflection "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/reflection.py"
click node_ingest_api "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/main.py"
click node_ingest_state "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/ingestion_state_manager.py"
click node_extraction "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/nvingest.py"
click node_schema "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/nemo_retriever/ingest_schema_manager.py"
click node_summarization "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/summarization.py"
click node_telemetry "https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/observability/tracing/instrumentation.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_ui,node_chat,node_collection_ui,node_settings_ui,node_user,node_api_client toneBlue
class node_ingest_api,node_ingest_state,node_extraction,node_schema,node_summarization toneAmber
class node_rag_api,node_query_processing,node_agentic,node_retrieval toneMint
class node_generation,node_vlm,node_reflection,node_safety toneRose
class node_vector_store,node_elasticsearch,node_milvus,node_nim_models,node_telemetry,node_documents toneIndigo
```

## Components

### User Experience

- [Reference UI](https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/App.tsx): `frontend/src/App.tsx`
- [Chat workflow](https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/pages/Chat.tsx): `frontend/src/pages/Chat.tsx`
- [Collection management](https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/pages/NewCollection.tsx): `frontend/src/pages/NewCollection.tsx`
- [Settings](https://github.com/nvidia-ai-blueprints/rag/blob/main/frontend/src/pages/SettingsPage.tsx): `frontend/src/pages/SettingsPage.tsx`

### Ingestion

- [Ingestion API](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/main.py): `src/nvidia_rag/ingestor_server/main.py`
- [Ingestion task state](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/ingestion_state_manager.py): `src/nvidia_rag/ingestor_server/ingestion_state_manager.py`
- [Document extraction](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/nvingest.py): `src/nvidia_rag/ingestor_server/nvingest.py`
- [Ingest schema adapter](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/ingestor_server/nemo_retriever/ingest_schema_manager.py): `src/nvidia_rag/ingestor_server/nemo_retriever/ingest_schema_manager.py`
- [Document summaries](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/summarization.py): `src/nvidia_rag/utils/summarization.py`

### Query and Retrieval

- [RAG API](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/main.py): `src/nvidia_rag/rag_server/main.py`
- [Query processing](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/query_decomposition.py): `src/nvidia_rag/rag_server/query_decomposition.py`
- [Agentic RAG](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/agentic_rag/agentic_rag.py): `src/nvidia_rag/rag_server/agentic_rag/agentic_rag.py`
- [Retrieval and reranking](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/reranker.py): `src/nvidia_rag/utils/reranker.py`

### Answering

- [Response generation](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/response_generator.py): `src/nvidia_rag/rag_server/response_generator.py`
- [Multimodal answering](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/vlm.py): `src/nvidia_rag/rag_server/vlm.py`
- [Reflection](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/rag_server/reflection.py): `src/nvidia_rag/rag_server/reflection.py`
- **Optional guardrails**

### Storage and Operations

- [Vector store](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/vdb/vdb_base.py): `src/nvidia_rag/utils/vdb/vdb_base.py`
- [Elasticsearch adapter](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/vdb/elasticsearch/elastic_vdb.py): `src/nvidia_rag/utils/vdb/elasticsearch/elastic_vdb.py`
- [Milvus adapter](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/vdb/milvus/milvus_vdb.py): `src/nvidia_rag/utils/vdb/milvus/milvus_vdb.py`
- **NVIDIA NIM services**
- [Telemetry and tracing](https://github.com/nvidia-ai-blueprints/rag/blob/main/src/nvidia_rag/utils/observability/tracing/instrumentation.py): `src/nvidia_rag/utils/observability/tracing/instrumentation.py`

### Other components

- **User**
- **Enterprise documents**
- **API client**

## Connections

- User → Reference UI: uses
- Reference UI → Chat workflow: routes to
- Reference UI → Collection management: routes to
- Reference UI → Settings: routes to
- Chat workflow → RAG API: submits query
- API client → RAG API: calls API
- RAG API → Query processing: processes query
- Query processing → Retrieval and reranking: prepares search
- RAG API → Agentic RAG: dispatches agentic
- Agentic RAG → Retrieval and reranking: executes searches
- Retrieval and reranking → Vector store: searches
- Elasticsearch adapter → Vector store: implements
- Milvus adapter → Vector store: implements
- RAG API → Response generation: generates answer
- Response generation → NVIDIA NIM services: requests inference
- RAG API → Multimodal answering: dispatches multimodal
- Multimodal answering → NVIDIA NIM services: requests vision inference
- RAG API → Reflection: invokes reflection
- RAG API → Optional guardrails: filters content
- Enterprise documents → Ingestion API: uploads
- User → Collection management: configures collection
- Collection management → Ingestion API: creates collection
- Ingestion API → Ingestion task state: tracks progress
- Ingestion API → Document extraction: dispatches extraction
- Document extraction → NVIDIA NIM services: extracts content
- Document extraction → Ingest schema adapter: adapts records
- Ingest schema adapter → Vector store: writes records
- Ingestion API → Document summaries: requests summaries
- RAG API → Telemetry and tracing: emits traces
- Ingestion API → Telemetry and tracing: emits traces
