How pani62172/antinomai works
Antinomai is an investment-research chat platform: users submit questions, chart images, or portfolio details; the API invokes a LangGraph agent that routes requests to specialized research flows and returns a report (and, for portfolio analysis, structured holdings data). The React interface also supports voice transcription and a session-scoped research-document vault. Route coverage follows the README; some agent implementations and UI-to-vault requests were not fully excerpted, so those relationships are shown only where documented or visible.
Main entry points: agents.py Intent router · main.py FastAPI endpoints · useChat.ts Chat session · App.jsx Chat application
Components 26
Research interface
- Chat application
frontend/src/App.jsx - Chat and uploads
frontend/src/components/ChatInput.tsx - Chat session
frontend/src/useChat.ts - Conversation view
frontend/src/components/ChatWindow.tsx - Message rendering
frontend/src/components/MessageBubble.tsx - Portfolio table
frontend/src/components/PortfolioTable.tsx - Session controls
frontend/src/components/Sidebar.tsx - Vault panel
frontend/src/components/VaultPanel.jsx
API and state
- FastAPI endpoints
backend/main.py - Agent state and models
backend/state.py - Runtime configuration
backend/config.py
Research routes
- Intent router
backend/agents.py - Debate panel
backend/agents.py - Chart analysis
backend/agents.py - Quick lookups
backend/agents.py - Fundamental analysis
backend/agents.py - Portfolio analysis
backend/agents.py - Vault research
backend/agents.py - Market and search tools
backend/tools.py
Research vault
- Document ingestion and retrieval
backend/vault.py - Vector store
Other components
- Researcher
- Groq models
- Tavily search
- Market data providers
- Groq transcription
Connections 34
- Researcher → Chat and uploads: enters query
frontend/src/components/ChatInput.tsx - Chat application → Chat session: uses state
frontend/src/App.jsx - Chat application → Chat and uploads: renders
frontend/src/App.jsx - Chat application → Conversation view: renders
frontend/src/App.jsx - Chat application → Session controls: renders
frontend/src/App.jsx - Chat and uploads → Chat session: sends query
frontend/src/App.jsx - Chat session → FastAPI endpoints: calls chat and upload
frontend/src/useChat.ts - Chat session → FastAPI endpoints: checks health and clears
frontend/src/useChat.ts - Chat and uploads → Groq transcription: requests transcription
frontend/src/components/ChatInput.tsx - Conversation view → Message rendering: renders messages
frontend/src/components/ChatWindow.tsx - Conversation view → Portfolio table: renders portfolio
frontend/src/components/ChatWindow.tsx - Session controls → Chat session: controls session
frontend/src/App.jsx - Vault panel → FastAPI endpoints: uses vault endpoints
- FastAPI endpoints → Intent router: invokes graph
backend/main.py - FastAPI endpoints → Document ingestion and retrieval: ingests and clears documents
backend/main.py - Intent router → Agent state and models: uses schemas
backend/agents.py - Intent router → Debate panel: routes to
README.md - Intent router → Chart analysis: routes to
README.md - Intent router → Quick lookups: routes to
README.md - Intent router → Fundamental analysis: routes to
README.md - Intent router → Portfolio analysis: routes to
README.md - Intent router → Vault research: routes to
backend/agents.py - Debate panel → Market and search tools: uses research tools
backend/agents.py - Quick lookups → Market and search tools: uses lookup tools
README.md - Fundamental analysis → Market data providers: fetches company data
backend/agents.py - Portfolio analysis → Market data providers: fetches prices and beta
backend/agents.py - Market and search tools → Market data providers: queries market data
README.md - Market and search tools → Tavily search: searches web
README.md - Debate panel → Groq models: generates theses
README.md - Chart analysis → Groq models: analyzes chart
README.md - Vault research → Document ingestion and retrieval: retrieves context
backend/agents.py - Document ingestion and retrieval → Vector store: stores and retrieves chunks
backend/vault.py - FastAPI endpoints → Chat session: returns SSE events
frontend/src/useChat.ts - FastAPI endpoints → Portfolio table: returns portfolio data
backend/main.py