# intel/enterprise-rag architecture

> Architecture diagram of the GitHub repository intel/enterprise-rag, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-07). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/intel/enterprise-rag
- Repository: https://github.com/intel/enterprise-rag

## Overview

Intel AI for Enterprise RAG is a platform layer for governed document ingestion and RAG workloads, including ChatQnA, DocSum, AudioQnA, and translation. Users ingest files or links through the admin UI, then ask questions in a chat UI; the composed pipeline embeds the query, retrieves and reranks documents, applies guardrails, builds a prompt, and calls the LLM. The README documents this flow and the platform integrations. Source excerpts confirm selected service and EDP behavior; links for unsampled subsystems follow README statements rather than inspected implementation.

Main entry points: [`src/gmc/cmd/router/main.go`](https://github.com/intel/enterprise-rag/blob/main/src/gmc/cmd/router/main.go) (Pipeline router), [`src/edp/app/tasks.py`](https://github.com/intel/enterprise-rag/blob/main/src/edp/app/tasks.py) (Enhanced data prep), [`src/ui/apps/chatqna/src/app/routes/chat/ChatConversationRoute.tsx`](https://github.com/intel/enterprise-rag/blob/main/src/ui/apps/chatqna/src/app/routes/chat/ChatConversationRoute.tsx) (ChatQnA UI), [`src/ui/packages/data-ingestion/src/components/DataIngestionTab/DataIngestionTab.tsx`](https://github.com/intel/enterprise-rag/blob/main/src/ui/packages/data-ingestion/src/components/DataIngestionTab/DataIngestionTab.tsx) (Ingestion UI)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_experience["User applications"]
  node_chat_ui["ChatQnA UI"]
  node_admin_ui["Ingestion UI"]
  node_docsum_ui["DocSum UI<br/>[DocSumRoute.tsx]"]
  node_audio_ui["AudioQnA UI"]
end

subgraph group_control["Pipeline control"]
  node_composer["Pipeline composer"]
  node_gmconnector["GMConnector resource"]
  node_operator["GMC operator"]
  node_router["Pipeline router<br/>[main.go]"]
end

subgraph group_ingestion["Document ingestion"]
  node_edp["Enhanced data prep<br/>[tasks.py]"]
  node_extract["Text extraction"]
  node_split["Text splitting"]
  node_ingest["Vector ingestion"]
end

subgraph group_rag["RAG services"]
  node_embed["Embedding service"]
  node_retrieve["Retriever"]
  node_rerank["Reranking service"]
  node_guardrails["Guardrail services"]
  node_prompt["Prompt template"]
  node_llm["LLM service"]
  node_history[("Chat history")]
  node_docsum["Document summarization"]
  node_audio_services["Speech services"]
  node_translation["Translation workload"]
end

subgraph group_platform["Shared platform"]
  node_service_core["Microservice framework<br/>[micro_service.py]"]
  node_doc_protocol["Document protocols<br/>[docarray.py]"]
end

node_user(("Enterprise user"))
node_gateway{{"Platform gateway"}}
node_keycloak{{"Keycloak identity"}}
node_object_store[("Object storage")]
node_vector_db[("Vector database")]
node_inference{{"Platform inference"}}

node_user -->|"asks questions"| node_chat_ui
node_user -->|"uploads documents"| node_admin_ui
node_user -->|"submits documents"| node_docsum_ui
node_user -->|"asks by voice"| node_audio_ui
node_chat_ui -->|"sends requests"| node_gateway
node_admin_ui -->|"starts ingestion"| node_edp
node_docsum_ui -.->|"submits summaries"| node_gateway
node_audio_ui -.->|"sends audio"| node_gateway
node_gateway -->|"authenticates users"| node_keycloak
node_gateway -->|"routes requests"| node_router
node_composer -->|"renders flow"| node_gmconnector
node_gmconnector -->|"reconciles workload"| node_operator
node_operator -->|"manages services"| node_router
node_edp -->|"reads documents"| node_object_store
node_edp -->|"extracts content"| node_extract
node_edp -->|"splits documents"| node_split
node_edp -->|"requests embeddings"| node_embed
node_edp -->|"submits chunks"| node_ingest
node_ingest -->|"writes vectors"| node_vector_db
node_router -->|"embeds query"| node_embed
node_router -->|"retrieves candidates"| node_retrieve
node_retrieve -->|"searches vectors"| node_vector_db
node_router -->|"reranks candidates"| node_rerank
node_router -->|"filters input"| node_guardrails
node_router -->|"builds prompt"| node_prompt
node_router -->|"requests completion"| node_llm
node_llm -->|"requests inference"| node_inference
node_router -->|"streams answer"| node_chat_ui
node_chat_ui -.->|"manages conversations"| node_history
node_router -.->|"runs summary flow"| node_docsum
node_router -.->|"runs voice flow"| node_audio_services
node_router -.->|"runs translation flow"| node_translation
node_service_core -->|"serves typed payloads"| node_doc_protocol

click node_chat_ui "https://github.com/intel/enterprise-rag/blob/main/src/ui/apps/chatqna/src/app/routes/chat/ChatConversationRoute.tsx"
click node_admin_ui "https://github.com/intel/enterprise-rag/blob/main/src/ui/packages/data-ingestion/src/components/DataIngestionTab/DataIngestionTab.tsx"
click node_docsum_ui "https://github.com/intel/enterprise-rag/blob/main/src/ui/apps/docsum/src/app/routes/docsum/DocSumRoute.tsx"
click node_audio_ui "https://github.com/intel/enterprise-rag/blob/main/src/ui/apps/audioqna/src/app/routes/chat/ChatConversationRoute.tsx"
click node_composer "https://github.com/intel/enterprise-rag/blob/main/deployment/scripts/compose_pipeline.py"
click node_gmconnector "https://github.com/intel/enterprise-rag/blob/main/src/gmc/api/v1alpha3/gmconnector_types.go"
click node_operator "https://github.com/intel/enterprise-rag/blob/main/src/gmc/internal/controller/gmconnector_controller.go"
click node_router "https://github.com/intel/enterprise-rag/blob/main/src/gmc/cmd/router/main.go"
click node_edp "https://github.com/intel/enterprise-rag/blob/main/src/edp/app/tasks.py"
click node_extract "https://github.com/intel/enterprise-rag/blob/main/src/comps/text_extractor/erag_text_extractor_microservice.py"
click node_split "https://github.com/intel/enterprise-rag/blob/main/src/comps/text_splitter/erag_text_splitter_microservice.py"
click node_ingest "https://github.com/intel/enterprise-rag/blob/main/src/comps/ingestion/erag_ingestion_microservice.py"
click node_embed "https://github.com/intel/enterprise-rag/blob/main/src/comps/embeddings/erag_embedding_microservice.py"
click node_retrieve "https://github.com/intel/enterprise-rag/blob/main/src/comps/retrievers/erag_retriever_microservice.py"
click node_rerank "https://github.com/intel/enterprise-rag/blob/main/src/comps/reranks/erag_reranking_microservice.py"
click node_guardrails "https://github.com/intel/enterprise-rag/tree/main/src/comps/guardrails"
click node_prompt "https://github.com/intel/enterprise-rag/blob/main/src/comps/prompt_template/erag_prompt_template_microservice.py"
click node_llm "https://github.com/intel/enterprise-rag/blob/main/src/comps/llms/erag_llm_microservice.py"
click node_history "https://github.com/intel/enterprise-rag/blob/main/src/comps/chat_history/erag_chat_history_microservice.py"
click node_docsum "https://github.com/intel/enterprise-rag/blob/main/src/comps/docsum/erag_docsum_microservice.py"
click node_audio_services "https://github.com/intel/enterprise-rag/blob/main/src/comps/asr/erag_asr_microservice.py"
click node_translation "https://github.com/intel/enterprise-rag/tree/main/src/comps"
click node_service_core "https://github.com/intel/enterprise-rag/blob/main/src/comps/cores/mega/micro_service.py"
click node_doc_protocol "https://github.com/intel/enterprise-rag/blob/main/src/comps/cores/proto/docarray.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_chat_ui,node_admin_ui,node_docsum_ui,node_audio_ui,node_user toneBlue
class node_composer,node_gmconnector,node_operator,node_router,node_object_store,node_vector_db toneAmber
class node_edp,node_extract,node_split,node_ingest toneMint
class node_embed,node_retrieve,node_rerank,node_guardrails,node_prompt,node_llm,node_history,node_docsum,node_audio_services,node_translation toneRose
class node_service_core,node_doc_protocol,node_gateway,node_keycloak,node_inference toneIndigo
```

## Components

### User applications

- [ChatQnA UI](https://github.com/intel/enterprise-rag/blob/main/src/ui/apps/chatqna/src/app/routes/chat/ChatConversationRoute.tsx): `src/ui/apps/chatqna/src/app/routes/chat/ChatConversationRoute.tsx`
- [Ingestion UI](https://github.com/intel/enterprise-rag/blob/main/src/ui/packages/data-ingestion/src/components/DataIngestionTab/DataIngestionTab.tsx): `src/ui/packages/data-ingestion/src/components/DataIngestionTab/DataIngestionTab.tsx`
- [DocSum UI](https://github.com/intel/enterprise-rag/blob/main/src/ui/apps/docsum/src/app/routes/docsum/DocSumRoute.tsx): `src/ui/apps/docsum/src/app/routes/docsum/DocSumRoute.tsx`
- [AudioQnA UI](https://github.com/intel/enterprise-rag/blob/main/src/ui/apps/audioqna/src/app/routes/chat/ChatConversationRoute.tsx): `src/ui/apps/audioqna/src/app/routes/chat/ChatConversationRoute.tsx`

### Pipeline control

- [Pipeline composer](https://github.com/intel/enterprise-rag/blob/main/deployment/scripts/compose_pipeline.py): `deployment/scripts/compose_pipeline.py`
- [GMConnector resource](https://github.com/intel/enterprise-rag/blob/main/src/gmc/api/v1alpha3/gmconnector_types.go): `src/gmc/api/v1alpha3/gmconnector_types.go`
- [GMC operator](https://github.com/intel/enterprise-rag/blob/main/src/gmc/internal/controller/gmconnector_controller.go): `src/gmc/internal/controller/gmconnector_controller.go`
- [Pipeline router](https://github.com/intel/enterprise-rag/blob/main/src/gmc/cmd/router/main.go): `src/gmc/cmd/router/main.go`

### Document ingestion

- [Enhanced data prep](https://github.com/intel/enterprise-rag/blob/main/src/edp/app/tasks.py): `src/edp/app/tasks.py`
- [Text extraction](https://github.com/intel/enterprise-rag/blob/main/src/comps/text_extractor/erag_text_extractor_microservice.py): `src/comps/text_extractor/erag_text_extractor_microservice.py`
- [Text splitting](https://github.com/intel/enterprise-rag/blob/main/src/comps/text_splitter/erag_text_splitter_microservice.py): `src/comps/text_splitter/erag_text_splitter_microservice.py`
- [Vector ingestion](https://github.com/intel/enterprise-rag/blob/main/src/comps/ingestion/erag_ingestion_microservice.py): `src/comps/ingestion/erag_ingestion_microservice.py`

### RAG services

- [Embedding service](https://github.com/intel/enterprise-rag/blob/main/src/comps/embeddings/erag_embedding_microservice.py): `src/comps/embeddings/erag_embedding_microservice.py`
- [Retriever](https://github.com/intel/enterprise-rag/blob/main/src/comps/retrievers/erag_retriever_microservice.py): `src/comps/retrievers/erag_retriever_microservice.py`
- [Reranking service](https://github.com/intel/enterprise-rag/blob/main/src/comps/reranks/erag_reranking_microservice.py): `src/comps/reranks/erag_reranking_microservice.py`
- [Guardrail services](https://github.com/intel/enterprise-rag/tree/main/src/comps/guardrails): `src/comps/guardrails`
- [Prompt template](https://github.com/intel/enterprise-rag/blob/main/src/comps/prompt_template/erag_prompt_template_microservice.py): `src/comps/prompt_template/erag_prompt_template_microservice.py`
- [LLM service](https://github.com/intel/enterprise-rag/blob/main/src/comps/llms/erag_llm_microservice.py): `src/comps/llms/erag_llm_microservice.py`
- [Chat history](https://github.com/intel/enterprise-rag/blob/main/src/comps/chat_history/erag_chat_history_microservice.py): `src/comps/chat_history/erag_chat_history_microservice.py`
- [Document summarization](https://github.com/intel/enterprise-rag/blob/main/src/comps/docsum/erag_docsum_microservice.py): `src/comps/docsum/erag_docsum_microservice.py`
- [Speech services](https://github.com/intel/enterprise-rag/blob/main/src/comps/asr/erag_asr_microservice.py): `src/comps/asr/erag_asr_microservice.py`
- [Translation workload](https://github.com/intel/enterprise-rag/tree/main/src/comps): `src/comps`

### Shared platform

- [Microservice framework](https://github.com/intel/enterprise-rag/blob/main/src/comps/cores/mega/micro_service.py): `src/comps/cores/mega/micro_service.py`
- [Document protocols](https://github.com/intel/enterprise-rag/blob/main/src/comps/cores/proto/docarray.py): `src/comps/cores/proto/docarray.py`

### Other components

- **Enterprise user**
- **Platform gateway**
- **Keycloak identity**
- **Object storage**
- **Vector database**
- **Platform inference**

## Connections

- Enterprise user → ChatQnA UI: asks questions (evidence: `README.md`)
- Enterprise user → Ingestion UI: uploads documents (evidence: `README.md`)
- Enterprise user → DocSum UI: submits documents (evidence: `README.md`)
- Enterprise user → AudioQnA UI: asks by voice (evidence: `README.md`)
- ChatQnA UI → Platform gateway: sends requests (evidence: `README.md`)
- Ingestion UI → Enhanced data prep: starts ingestion (evidence: `README.md`)
- DocSum UI → Platform gateway: submits summaries
- AudioQnA UI → Platform gateway: sends audio
- Platform gateway → Keycloak identity: authenticates users (evidence: `README.md`)
- Platform gateway → Pipeline router: routes requests (evidence: `README.md`)
- Pipeline composer → GMConnector resource: renders flow (evidence: `README.md`)
- GMConnector resource → GMC operator: reconciles workload (evidence: `README.md`)
- GMC operator → Pipeline router: manages services (evidence: `README.md`)
- Enhanced data prep → Object storage: reads documents (evidence: `README.md`)
- Enhanced data prep → Text extraction: extracts content (evidence: `README.md`)
- Enhanced data prep → Text splitting: splits documents (evidence: `README.md`)
- Enhanced data prep → Embedding service: requests embeddings (evidence: `README.md`)
- Enhanced data prep → Vector ingestion: submits chunks (evidence: `src/edp/app/tasks.py`)
- Vector ingestion → Vector database: writes vectors (evidence: `README.md`)
- Pipeline router → Embedding service: embeds query (evidence: `README.md`)
- Pipeline router → Retriever: retrieves candidates (evidence: `README.md`)
- Retriever → Vector database: searches vectors (evidence: `README.md`)
- Pipeline router → Reranking service: reranks candidates (evidence: `README.md`)
- Pipeline router → Guardrail services: filters input (evidence: `README.md`)
- Pipeline router → Prompt template: builds prompt (evidence: `README.md`)
- Pipeline router → LLM service: requests completion (evidence: `README.md`)
- LLM service → Platform inference: requests inference (evidence: `README.md`)
- Pipeline router → ChatQnA UI: streams answer (evidence: `README.md`)
- ChatQnA UI → Chat history: manages conversations
- Pipeline router → Document summarization: runs summary flow (evidence: `README.md`)
- Pipeline router → Speech services: runs voice flow (evidence: `README.md`)
- Pipeline router → Translation workload: runs translation flow (evidence: `README.md`)
- Microservice framework → Document protocols: serves typed payloads (evidence: `src/comps/cores/mega/micro_service.py`)
