Tagged
14 posts · page 1 of 2
Apple and Google both want assistants to act inside your app, and they picked completely different ways to do it: Swift App Intents on one side, Kotlin AppFunctions on the other. I didn't want to write the same feature twice in two languages I don't ship in, so I built a source generator that writes both from one C# declaration.
v13 puts AES-256-GCM field encryption in the core package, adds a transactional outbox that commits with the write that caused it, and gives a document type a front door — REST + Server-Sent Events, and an MCP server. Plus Orleans streams on the database you already have, and admin tools that read an encrypted store without a key.
A MAUI UI July post: two drop-in MAUI controls — FaceRecognitionView and FaceEnrollmentView — that put live face recognition and a guided enrollment wizard on screen in one line of XAML, plus the CameraView frame analyzer, a GraphicsView overlay with an even-odd scrim, speaker recognition, and Shiny.DocumentDb + sqlite-vec as the on-device vector database.
A .NET MAUI app built straight on Microsoft.Extensions.AI — an IChatClient onto OpenAI, ChatOptions.Tools, and UseFunctionInvocation running the loop. The chat UI is a control; the hands are Shiny — contacts, reminders, calendar, GPS, health, music, a local document DB, your own services, and the app's own navigation. All opt-in, all switchable, all AOT-clean.
Three new Microsoft.Extensions.AI packages turn device services into tools an LLM can call — Shiny.Contacts.Extensions.AI, Shiny.Notifications.Extensions.AI, and Shiny.Locations.Extensions.AI. Opt-in, read-only by default, AOT-safe, and they all compose into one IChatClient.
One IHealthService for Apple HealthKit and Android Health Connect — now with 30+ data types across activity, body, vitals, nutrition, reproductive/cycle tracking and workouts, real-time observation, and a Microsoft.Extensions.AI tool surface so an LLM can read and write your health data.
v6 adds cross-provider vector / ANN search, EF-style global query filters, composite JSON indexes, real connection pooling on server SQL, and per-query change monitoring — without changing the schema-free, AOT-safe shape of the API.
Shiny.AiConversation wraps chat clients, speech recognition, text-to-speech, audio playback, message persistence, and state management into a single IAiConversationService interface for .NET.
Shiny.DocumentDb.Extensions.AI turns your document store into a set of LLM-callable tools. Register types, set capabilities, and let the agent query, insert, update, and delete documents through natural language.
Mark a service interface with [Tool], add [Description] to your methods, and a source generator produces fully typed AIFunction classes compatible with Microsoft.Extensions.AI. AOT-safe, zero boilerplate.
Write a single C# record and have it automatically become a fully typed AI tool — with zero adapter code. Shiny Mediator 6.3 delivers contract-first source generation for AI tools, HTTP clients, and ASP.NET endpoints.
How Shiny MAUI Shell uses source generation and Microsoft.Extensions.AI to let an AI chat agent discover your app routes, match user intent, and navigate with pre-filled forms.
