How nextgennvidia/riva-agi works
Riva-AGI combines multi-agent task orchestration, a document-backed RAG knowledge service, and a Gemini-powered voice interface. The principal task workflow accepts a request, classifies and routes it through orchestration, invokes a registered agent and its tools, then returns a response. RAG separately ingests and privacy-processes documents for retrieval and answer synthesis; voice provides a browser-facing conversation service. Relationships are limited to those shown in excerpts or the source index; unsampled integrations are omitted where their wiring is uncertain.
Main entry points: main.py Orchestrator graph · web_server.py Voice web server · session.py Gemini voice session · pipeline.py Document ingestion
Components 32
Task orchestration
- Chat entry
chat.py - Orchestrator graph
orchestration/orchestrator/main.py - Intent routing
orchestration/orchestrator/intent_classifier.py - Task planner
orchestration/orchestrator/planner.py - Task executor
orchestration/orchestrator/executor.py - Response reviewer
orchestration/orchestrator/reviewer.py - Task state
orchestration/orchestrator/state_manager.py
Agents and tools
- Agent registry
orchestration/orchestrator/registry.py - Specialist agents
orchestration/agents/researcher.py - Gemini agent runtime
orchestration/orchestrator/llm.py - Tool registry
orchestration/tools/registry.py - Built-in tools
orchestration/tools/builtin/web_tools.py
RAG knowledge
- Document ingestion
rag_knowledge/ingestion/pipeline.py - Document loaders
rag_knowledge/ingestion/loaders/dispatch.py - Privacy checks
rag_knowledge/ingestion/privacy.py - PDF reader
rag_knowledge/ingestion/loaders/pdf.py - Word reader
rag_knowledge/ingestion/readers/word.py - Qdrant store
rag_knowledge/storage/qdrant_storage.py - Knowledge retrieval
rag_knowledge/retrieval/retriever.py - RAG answer service
rag_knowledge/service.py - Gemini synthesis
rag_knowledge/clients/gemini_client.py - RAG prompts
rag_knowledge/prompts.py
Voice conversation
- Voice web server
voice_speech/web_server.py - Gemini voice session
voice_speech/engine/gemini/session.py - Voice streaming
voice_speech/engine/gemini/streaming.py - Voice tools
voice_speech/engine/gemini/tools.py - Conversation state
voice_speech/engine/conversation/state.py - Session manager
voice_speech/engine/conversation/session_manager.py - Voice settings
voice_speech/engine/config/settings.py - Voice prompts
voice_speech/engine/config/prompts.py
Other components
- User
- Browser user
Connections 31
- User → Chat entry: submits task
chat.py - Chat entry → Orchestrator graph: runs task
chat.py - Orchestrator graph → Intent routing: classifies request
orchestration/orchestrator/main.py - Orchestrator graph → Task planner: plans work
orchestration/orchestrator/main.py - Orchestrator graph → Task executor: dispatches work
orchestration/orchestrator/main.py - Orchestrator graph → Response reviewer: routes review
orchestration/orchestrator/main.py - Orchestrator graph → Agent registry: looks up agents
orchestration/orchestrator/main.py - Orchestrator graph → Task state: updates task state
orchestration/orchestrator/main.py - Specialist agents → Agent registry: registers agents
orchestration/agents/researcher.py - Orchestrator graph → Specialist agents: invokes handler
orchestration/orchestrator/main.py - Specialist agents → Gemini agent runtime: calls Gemini
orchestration/agents/researcher.py - Gemini agent runtime → Tool registry: resolves tools
orchestration/orchestrator/llm.py - Tool registry → Built-in tools: registers tools
orchestration/tools/builtin/web_tools.py - Document ingestion → Document loaders: loads documents
rag_knowledge/ingestion/pipeline.py - Document ingestion → Privacy checks: checks privacy
rag_knowledge/ingestion/pipeline.py - Document loaders → PDF reader: dispatches PDFs
rag_knowledge/ingestion/loaders/dispatch.py - Document loaders → Word reader: dispatches Word files
rag_knowledge/ingestion/loaders/dispatch.py - Document ingestion → Qdrant store: upserts documents
rag_knowledge/ingestion/pipeline.py - RAG answer service → Knowledge retrieval: retrieves matches
rag_knowledge/service.py - Knowledge retrieval → Qdrant store: searches store
rag_knowledge/retrieval/retriever.py - RAG answer service → Gemini synthesis: synthesizes answer
rag_knowledge/service.py - Gemini synthesis → RAG prompts: formats prompt
rag_knowledge/clients/gemini_client.py - Browser user → Voice web server: opens conversation
- Voice web server → Session manager: manages sessions
voice_speech/web_server.py - Voice web server → Conversation state: uses state
voice_speech/web_server.py - Voice web server → Gemini voice session: starts sessions
voice_speech/web_server.py - Voice web server → Voice streaming: uses streaming
voice_speech/web_server.py - Gemini voice session → Voice settings: loads settings
voice_speech/engine/gemini/session.py - Gemini voice session → Voice prompts: loads prompts
voice_speech/engine/gemini/session.py - Gemini voice session → Voice tools: uses tools
voice_speech/engine/gemini/session.py - Voice tools → Tool registry: uses agent tools
voice_speech/engine/gemini/tools.py