Orchestrate specialized AI agents across a software-development lifecycle with human-in-the-loop checkpoints, persistent memory, and MCP tool integration.
| Doc | Description |
|---|---|
| Architecture | System design, component diagram, data flow |
| API Reference | REST endpoints, request/response schemas |
| Configuration | Environment variables, appsettings, secrets |
| Security | RBAC, GDPR, EU AI Act, pen-test fixes |
| Development Setup | Prerequisites, docker compose, run locally |
| Contributing | Conventions, branching, testing |
| Layer | Technology |
|---|---|
| Orchestration runtime | Microsoft Orleans 10 (ADO.NET / PostgreSQL) |
| API | ASP.NET Core 10 (REST + SignalR + MCP SSE) |
| LLM abstraction | Microsoft.Extensions.AI (IChatClient) |
| MCP | ModelContextProtocol 1.1 (official SDK) |
| Vector DB | Qdrant (Docker) |
| Relational DB | PostgreSQL 16 (EF Core 10) |
| Pub/Sub + KV | Redis 7 |
| Dashboard | Blazor Server |
| Auth | Keycloak 26 (OAuth2/OIDC, JWT Bearer) |
| Observability | Serilog → Seq |
cp infra/.env.example infra/.env # fill in real secrets
cd infra && docker compose up -d
dotnet run --project src/Nox.Api
# browse http://localhost:5000/swaggerSee dev-setup.md for details.