How uday-sr/second-brain works
Second Brain is a user-facing knowledge library: users sign in, save content, and explore saved items; supported sources can then be indexed for contextual Q&A. The React app talks to the Node API for accounts and content, while the Python AI service exposes separate ingestion and question-answering endpoints. Its ingestion path loads a source, chunks and embeds it, then stores an index; Q&A retrieves context and asks Ollama for an answer. The README describes semantic search and summaries, though the sampled AI endpoint shows Q&A rather than a separate summary flow. Some frontend-to-AI and backend-to-AI wiring is not shown, so those connections are omitted.
Main entry points: main.py AI service API · api.ts API clients · Home.tsx Content home · Card.tsx Content cards
Components 25
Web application
- React app
frontend/src/main.tsx - Intro and auth
frontend/src/routes/Intro.tsx - Sign-in modal
frontend/src/components/AuthModal.tsx - Content home
frontend/src/routes/Home.tsx - Add content
frontend/src/components/ContentModal.tsx - Content cards
frontend/src/components/Card.tsx - Ask AI UI
frontend/src/components/AskAi.tsx - API clients
frontend/src/lib/api.ts
Content API
- Express app
backend/src/server/app.ts - User routes
backend/src/Routers/user.ts - Content routes
backend/src/Routers/content.ts - User middleware
backend/src/middlewares/user.ts
AI ingestion
- AI service API
ai_service/app/main.py - Source loader registry
ai_service/app/loaders/__init__.py - YouTube loader
ai_service/app/loaders/youtube.py - Transcript extraction
ai_service/app/yt/downloader.py - Text chunker
ai_service/app/rag/chunker.py - Embedding model
ai_service/app/rag/embedder.py
AI retrieval and Q&A
- Vector retrieval
ai_service/app/rag/retriever.py - Answer generation
ai_service/app/rag/llm.py
Persistence and identity
- Prisma client
backend/src/lib/prisma.ts - PostgreSQL
- Index and status store
ai_service/app/store.py
Other components
- User
- Ollama
Connections 28
- User → React app: uses
- React app → Intro and auth: renders
- Intro and auth → Sign-in modal: opens
frontend/src/routes/Intro.tsx - Sign-in modal → API clients: submits credentials
frontend/src/components/AuthModal.tsx - Content home → API clients: fetches content
frontend/src/routes/Home.tsx - Content home → Add content: opens
frontend/src/routes/Home.tsx - Content home → Content cards: renders
frontend/src/routes/Home.tsx - Content cards → Ask AI UI: renders
frontend/src/components/Card.tsx - Content cards → API clients: calls
frontend/src/components/Card.tsx - Add content → API clients: creates content
frontend/src/components/ContentModal.tsx - API clients → Express app: requests
frontend/src/lib/api.ts - Express app → User routes: registers route
backend/src/server/app.ts - Express app → Content routes: registers route
backend/src/server/app.ts - User routes → Prisma client: queries
backend/src/Routers/user.ts - Content routes → Prisma client: queries
backend/src/Routers/content.ts - User routes → User middleware: uses
backend/src/Routers/user.ts - Content routes → User middleware: uses
backend/src/Routers/content.ts - Prisma client → PostgreSQL: persists data
- AI service API → Source loader registry: loads source
ai_service/app/main.py - Source loader registry → YouTube loader: dispatches
ai_service/app/loaders/__init__.py - YouTube loader → Transcript extraction: extracts transcript
ai_service/app/loaders/youtube.py - AI service API → Text chunker: chunks segments
ai_service/app/main.py - AI service API → Embedding model: embeds chunks
ai_service/app/main.py - AI service API → Index and status store: saves index
ai_service/app/main.py - AI service API → Index and status store: reads status and index
ai_service/app/main.py - AI service API → Vector retrieval: retrieves matches
ai_service/app/main.py - AI service API → Answer generation: requests answer
ai_service/app/main.py - Answer generation → Ollama: generates response
ai_service/app/rag/llm.py