NER System Architecture

NER System Architecture An architecture diagram generated by Archify. Widget Tree · Mobile + Desktop UI · Flutter App Widget Tree Mobile + Desktop UI State Mgmt · BLoC / Riverpod · Flutter App State Mgmt BLoC / Riverpod Network Service · Dio + auth interceptor · Flutter App Network Service Dio + auth interceptor Local Storage · Prefs, Hive, Device APIs · Flutter App Local Storage Prefs, Hive, Device APIs UI / View Layer · Pages, Components · Website UI / View Layer Pages, Components State & Router · Auth, Store, Routes · Website State & Router Auth, Store, Routes API Client · HTTP, Interceptors · Website · response cache API Client HTTP, Interceptors response cache Routing & Proxy · Handlers, Middleware · API Gateway / Backend Routing & Proxy Handlers, Middleware Auth & Security · JWT, Rate limit, CORS · API Gateway / Backend Auth & Security JWT, Rate limit, CORS Controllers · App handler, ML proxy · API Gateway / Backend Controllers App handler, ML proxy Data Access · ORM, Conn pooling · API Gateway / Backend Data Access ORM, Conn pooling Inference API · FastAPI / Triton · Shared ML Model Service Inference API FastAPI / Triton Preprocessing · Cleansing, Features · Shared ML Model Service Preprocessing Cleansing, Features Model Runtime · ONNX / PyTorch · Shared ML Model Service Model Runtime ONNX / PyTorch Inference Cache · Outputs, Features · Shared ML Model Service Inference Cache Outputs, Features Pool & Executor · Connections, Queries · Shared Database Pool & Executor Connections, Queries DB Engine · Tables, Indexes · Shared Database DB Engine Tables, Indexes Shared Schemas · Users, State, ML logs · Shared Database Shared Schemas Users, State, ML logs persist HTTPS / JSON HTTPS / JSON ML proxy SQL lookup log predictions Flutter App Website API Gateway / Backend Shared ML Model Service Shared Database

Clients

  • • Flutter: widgets → BLoC/Riverpod → Dio
  • • Website: pages → store/router → API client

Gateway

  • • Route → JWT/rate limit/CORS → controllers
  • • Controllers call the DB layer or proxy to ML

ML + Data

  • • Cache hit or preprocess → model runtime
  • • One pool, engine, and schema set