Multi-Platform Voice Assistant Development

I need a speech-driven virtual assistant that lives on iOS, Android, and in a lightweight web client. Its core jobs are threefold: handle calendar scheduling with smart reminders, fetch facts or real-time information on request, and trigger common home-automation actions such as lights, thermostat, or security scenes.

I am looking for an end-to-end build that covers natural-language understanding, speech-to-text, text-to-speech, and a small but polished UI shell for each platform. You are free to choose the NLP stack—Dialogflow, Rasa, or a custom transformer model—so long as latency stays low and the solution can scale. For voice, feel free to tap into native engines (Apple Speech, Google Speech, Web Speech) or a cloud provider like Azure Cognitive Services if it speeds development.

Deliverables
• Cross-platform codebase with a shared intent engine and separate UI wrappers for iOS (Swift/SwiftUI), Android (Kotlin/Jetpack), and the web (React or similar)
• Calendar integration that can read, create, and update events in Google, Outlook, and Apple calendars
• Knowledge lookup module that hits at least one public API (e.g., Wikipedia, OpenWeather, or News) and returns concise answers in speech
• Home-automation bridge supporting common MQTT or Home Assistant endpoints, with room to add more via plugins
• A simple admin panel or config file that lets me map new intents to actions without redeploying code
• Short setup guide plus recorded demo showing the assistant scheduling a meeting, answering a fact query, and turning on a smart light

I’ll test on an iPhone, a recent Android handset, and Chrome; acceptance is based on smooth voice interaction, accurate intent recognition, and successful execution of all three task categories.