How heyito/ito works
Project type and purpose
• This is “Ito,” a cross-platform desktop voice-dictation app built on Electron.
• It combines a multi-process client (main, renderer, preload, native binaries), a gRPC transcription server (with a database and AWS infra), and native components in Rust/Swift.
File structure
• Top-level directories:
• app/ – Electron renderer (React + Zustand + Tailwind)
• lib/ – Shared TypeScript code for Electron main, preload, media, protocol, types, utils
• native/ – Rust & Swift binaries for audio capture, global key listening, text insertion, app context
• server/ – gRPC transcription server (Node.js + Bun + TypeScript), DB code, migrations, infra-as-code (AWS CDK)
• Key build/config files: electron.vite.config.ts, electron-builder.config.js, tsconfig.*, Dockerfiles, build-scripts
• Workflow: user interaction → Electron main → native binaries → gRPC client → server → AI model → back through main → native text writer → target app
Extract extra context from README
• Client architecture:
• Main process: system integration, permissions, orchestrates native components
• Renderer process: React UI (components, store, styles)
• Preload: secure IPC bridge
• Native components: Rust/Swift for performance
• Server:
• gRPC API, database (PostgreSQL via Bun/TypeORM/SQLite in dev), migrations, AWS infra (Fargate, Lambdas, Firehose, OpenSearch)
• Key external services: Auth0, AWS (Cognito, Firehose, OpenSearch), LLM/A I transcription providers