Join AI in Making Epic Stories — Agent Framework Edition

JAIMES AF is a .NET Aspire solution that coordinates FastEndpoints APIs, Blazor UI, EF Core repositories, document-processing workers, and Azure OpenAI-powered assistants to help groups co-create roleplaying adventures with rich observability and testing.

• Getting Started - Setup, configuration, and development guidelines.
• Architecture Overview - High-level system architecture, design patterns, and data flow.
• Agent Framework - Concepts behind the AI agent integration, tools, and conversation threads.
• Chat Streaming - Architecture of the real-time chat streaming system (Server-Sent Events).
• Message Bus & Workers - Overview of the message bus and background worker roles.
• Vector Search & RAG - Explanation of the RAG pipeline and vector search implementation.
• Observability - OpenTelemetry configuration, Aspire dashboard, and telemetry debugging.
• Database Schema - Database schema and entity relationships.
• Technical Topics - Comprehensive index of technical topics.
• UI Style Guide - UI conventions, reusable components, and styling patterns for the Blazor frontend.
• Pages - Comprehensive index of pages in the Blazor frontend with style guide compliance and improvement suggestions.

The Aspire AppHost wires together the API, background workers, and dependencies (PostgreSQL, Qdrant, Azure OpenAI). The diagram below shows the high-level structure.

flowchart LR
    subgraph Client Tier
        Blazor[JAIMES AF.Web\nBlazor Web UI]
    end
    subgraph Application Tier
        ApiService[JAIMES AF.ApiService\nFastEndpoints]
        Services[JAIMES AF.Services\nBusiness Logic]
        Repos[JAIMES AF.Repositories\nEF Core]
        Workers[JAIMES AF.Workers.*\nBackground Pipelines]
    end
    subgraph Data & Intelligence
        PostgreSQL[(PostgreSQL)]
        Qdrant[(Qdrant Vector DB)]
        AOAI[(Azure OpenAI / Ollama)]
    end

    Blazor -->|REST/gRPC| ApiService
    ApiService --> Services
    Services --> Repos
    Repos --> PostgreSQL
    Services --> Qdrant
    Services --> AOAI
    Workers --> Qdrant
    Workers --> Repos
    Workers --> AOAI

sequenceDiagram
    participant Player
    participant Web as Blazor UI
    participant API as FastEndpoints API
    participant BLL as Services & Mappers
    participant DB as EF Core / PostgreSQL
    participant Vector as Qdrant Vector DB
    participant AOAI as Azure OpenAI

    Player->>Web: Create Scenario / Send Message
    Web->>API: HTTP request with DTO payload
    API->>BLL: Validate + map to command
    BLL->>DB: Persist games, players, chat
    BLL->>Vector: Search rules / history
    BLL->>AOAI: Generate response
    AOAI-->>BLL: AI Completion
    Workers->>AOAI: Generate embeddings / responses
    Workers->>DB: Store chunk + trace metadata
    API-->>Web: Return updated DTO

• FastEndpoints-backed API surface for game, player, scenario, and ruleset management.
• Modular Service Layer with clear interfaces and business logic encapsulation.
• EF Core Repositories supporting PostgreSQL with robust migrations.
• Document Processing Pipeline that cracks, chunks, and embeds documents for RAG.
• Azure OpenAI Integration for intelligent, context-aware responses.
• Observability baked in via OpenTelemetry and Aspire dashboards.

For deep dives into these capabilities, please refer to the files listed in the Documentation Index.