application-engineering

What problem does it solve?

application-engineering turns AI-backed REST/API requirements into production-ready service boundaries with strong input/output validation, reliable Claude integration, and required telemetry.

Core Features & Use Cases

• FastAPI API design with Pydantic contracts: Build request/response models that produce correct OpenAPI docs and prevent internal-field leakage.
• Single-call Claude integration via an AICompletionPort: Use a provider-agnostic port with a Claude adapter to keep routers thin and tests deterministic using FakeAIPort.
• Mandatory OTel instrumentation at the request boundary: Emit one span per Claude call for latency and cost visibility.
• Auth, error handling, and streaming to clients: Apply authentication/authorization patterns, add backoff for rate limits, and support SSE streaming responses.

Quick Start

Use application-engineering to design a FastAPI endpoint that accepts a validated request, calls Claude through an AICompletionPort once per request, and returns a structured or streamed response with OTel spans for the Claude call.