FastAPI

Alguard的Tier 2是一个独立的FastAPI服务（web_api/resolve_action.py），运行在8010端口，拥有独立的OAuth audience和独立的approve scope，是唯一能解除HALTED状态的地方。

Somo's backend uses FastAPI plus Server-Sent Events for streaming output, and the frontend is a single-file mobile interface.

New versions of Flask have begun to support ASGI, and the performance gap between FastAPI and Flask may narrow in the future, but the advantages of asynchronous architecture in high-concurrency scenarios will still remain.

At startup, FastAPI scans the routes and type annotations in the code, automatically generates a description file conforming to the OpenAPI specification, and then hands it to Swagger UI or ReDoc to render into an interactive page where users can fill in parameters and initiate real requests for debugging.

FastAPI has built-in automatic documentation generation; after starting the application, accessing a specific path lets you view visual API documentation, including information such as the request parameters, request bodies, and response models of all endpoints.

FastAPI uses Python type annotations simultaneously for three things: parameter parsing, data validation, and documentation generation.

The advantages of asynchronous IO are mainly manifested in IO-intensive task scenarios, such as web interfaces waiting for database queries, downstream service responses, and file reads and writes; for CPU-intensive tasks, the benefit of asynchronous switching is limited, and they usually need to be placed into a thread pool or task queue for execution.

The official documentation lists FastAPI's key features as fast, efficient coding, fewer bugs, standards-based, and so on.

DecisionShield adopts a separated frontend-backend architecture: the frontend uses React+Vite+Tailwind CSS+Recharts+Framer Motion, the backend uses Python+FastAPI+Pydantic+Uvicorn, the AI layer uses 4-bit quantized local inference of Gemma 4 E4B, and the MVP stage uses in-memory storage.

VaultRAG's tech stack is: LLM and embedding models run locally via Ollama (using Llama 3.1, Qwen and nomic-embed-text), the backend uses Python and FastAPI, RAG orchestration uses LangChain, the frontend uses React with Tailwind CSS, and deployment is managed by Docker Compose.