FastAPI

Execution Workflow

1. Configure SQLAlchemyAsyncConfig and attach AdvancedAlchemy to the FastAPI app.
2. Build request-scoped session dependencies with Depends(alchemy.provide_session()).
3. Provide services through async generators and Service.new(session=...) context managers.
4. Keep routers thin and return service-converted schema objects or paginated schema collections.
5. Register routers only after the service and dependency layer is coherent.

Implementation Rules

• Use Annotated plus Depends for session and service injection to keep signatures explicit.
• Set commit_mode intentionally; do not rely on implicit transaction assumptions.
• Keep response_model aligned with the schema returned by to_schema().
• Avoid passing ORM instances directly through FastAPI responses.

Example Pattern

from advanced_alchemy.extensions.fastapi import AdvancedAlchemy, AsyncSessionConfig, SQLAlchemyAsyncConfig
from fastapi import FastAPI

alchemy = AdvancedAlchemy(
    config=SQLAlchemyAsyncConfig(
        connection_string="sqlite+aiosqlite:///test.sqlite",
        session_config=AsyncSessionConfig(expire_on_commit=False),
        create_all=True,
        commit_mode="autocommit",
    ),
    app=FastAPI(),
)

Validation Checklist

• Confirm the session dependency is request-scoped and closes after each request.
• Confirm service providers yield the expected service type and release resources.
• Confirm response models match the schema conversion output.
• Confirm routers are mounted only once and startup or shutdown lifecycle is not duplicated.

Cross-Skill Handoffs

• Use advanced-alchemy-routing for CRUD endpoint shape.
• Use advanced-alchemy-services for service-backed FastAPI handlers.
• Use advanced-alchemy-cli if the project needs Advanced Alchemy migration commands alongside the app.

Advanced Alchemy References

• https://advanced-alchemy.litestar.dev/latest/usage/frameworks/fastapi.html
• https://github.com/litestar-org/advanced-alchemy/blob/main/README.md