How letuandatt/ai-radar works
AI-Radar collects AI research and news, turns it into validated knowledge, and serves it through a persistent knowledge base, retrieval, and daily intelligence. The graph also includes the documented web reading experience and Zalo delivery. Several pipeline, digest, and intelligence implementations are unsampled, so those documented capabilities are shown at subsystem level; edges are limited to relationships visible in excerpts, the source index, or README.
Main entry points: routes.py Versioned API · application.py Lifecycle orchestration · retrieval_service.py Retrieval service · App.tsx Web application
Components 21
Application runtime
- Application entry
app/main.py - Lifecycle orchestration
app/core/application.py - Versioned API
app/api/v1/routes.py
Knowledge ingestion
- Source fetchers
app/fetchers - Acquisition pipeline
app/pipelines/acquisition.py - Processing pipeline
app/pipelines/processing.py - Manual URL add
app/services/knowledge/manual_add_service.py
Knowledge and retrieval
- Knowledge object
app/models/knowledge_object.py - Repository facade
app/services/app_service.py - Knowledge repository
app/storage/knowledge/sqlite_store.py - Vector synchronization
app/services/vector/sync_service.py - Qdrant vectors
app/storage/vector/qdrant_store.py - Retrieval service
app/services/retrieval/retrieval_service.py - Keyword index
app/storage/search/bm25_index.py
Intelligence and delivery
- Daily digest
app/pipelines/daily_digest.py - Analysis services
app/services/analysis - Zalo delivery
app/integrations/zalo
Web experience
- Web application
web/src/App.tsx - Web data hooks
web/src/hooks - Generated API client
web/src/api/generated.ts
Other components
- Reader
Connections 23
- Reader → Web application: browses
- Web application → Web data hooks: uses views
web/src/App.tsx - Web data hooks → Generated API client: calls client
web/src/hooks/useKnowledge.ts - Web application → Versioned API: reads stored content
- Application entry → Lifecycle orchestration: starts
app/main.py - Lifecycle orchestration → Acquisition pipeline: schedules
app/core/application.py - Lifecycle orchestration → Processing pipeline: forwards articles
app/core/application.py - Source fetchers → Acquisition pipeline: provides sources
- Acquisition pipeline → Processing pipeline: supplies articles
- Reader → Versioned API: submits URL
- Versioned API → Manual URL add: adds knowledge
- Processing pipeline → Knowledge object: produces knowledge
- Manual URL add → Knowledge object: builds object
app/services/knowledge/manual_add_service.py - Repository facade → Knowledge repository: persists records
app/services/app_service.py - Repository facade → Vector synchronization: synchronizes vectors
app/services/app_service.py - Vector synchronization → Qdrant vectors: upserts vectors
app/services/vector/sync_service.py - Versioned API → Retrieval service: dispatches search
app/api/v1/routes.py - Retrieval service → Keyword index: searches keywords
app/services/retrieval/retrieval_service.py - Retrieval service → Qdrant vectors: searches vectors
- Versioned API → Knowledge repository: browses records
- Knowledge repository → Daily digest: provides knowledge
- Daily digest → Zalo delivery: delivers digest
- Analysis services → Daily digest: informs intelligence