Recommended Stacks

Pre-tested package combinations for common Azure Functions project types.

Each stack lists the toolkit packages to install, what they provide together, and the scaffold command to generate a matching project.

API Stack

Best for: REST APIs, webhooks, and public-facing HTTP services.

Package	Role
azure-functions-openapi	Swagger UI + OpenAPI 3.1 spec
azure-functions-validation	Pydantic request/response validation
azure-functions-logging	Structured JSON logging
azure-functions-doctor	Pre-deploy diagnostic checks

afs new my-api

# requirements.txt
azure-functions
azure-functions-openapi
azure-functions-validation
azure-functions-logging

What you get: validated HTTP endpoints with auto-generated API docs, structured logs, request correlation IDs, and pre-wired azure-functions-doctor health checks.

---

DB API Stack

Best for: CRUD APIs backed by PostgreSQL, MySQL, or SQL Server.

Package	Role
azure-functions-openapi	Swagger UI + OpenAPI 3.1 spec
azure-functions-validation	Pydantic request/response validation
azure-functions-db	Declarative database input/output bindings
azure-functions-logging	Structured JSON logging

afs new my-api

# requirements.txt
azure-functions
azure-functions-openapi
azure-functions-validation
azure-functions-db[postgres]
azure-functions-logging

What you get: everything in the API Stack plus declarative database read/write bindings via @db.input() and @db.output().

Manual step: afs new does not yet ship a --with-db flag. After scaffolding, add azure-functions-db[postgres] (or another supported engine) to requirements.txt and wire up your bindings under app/.

---

AI Agent Stack

Best for: LangGraph-based conversational agents and AI workflows.

Package	Role
azure-functions-langgraph	LangGraph HTTP deployment adapter
azure-functions-openapi	Swagger UI for agent endpoints
azure-functions-logging	Structured JSON logging

afs ai agent my-agent

# requirements.txt
azure-functions
langgraph
azure-functions-langgraph
azure-functions-openapi
azure-functions-logging

What you get: invoke, stream, and health endpoints for LangGraph graphs, plus API documentation and structured logging.

---

Full Stack

Best for: production applications that need everything — validated APIs, database access, health diagnostics, and observability.

Package	Role
azure-functions-openapi	Swagger UI + OpenAPI 3.1 spec
azure-functions-validation	Pydantic request/response validation
azure-functions-db	Declarative database bindings
azure-functions-logging	Structured JSON logging
azure-functions-doctor	Pre-deploy diagnostic checks

afs new my-api

# requirements.txt
azure-functions
azure-functions-openapi
azure-functions-validation
azure-functions-db[postgres]
azure-functions-logging
azure-functions-doctor

What you get: a fully instrumented API with validation, database access, API docs, structured logging, and pre-wired health diagnostics.

Manual step: afs new does not yet ship a --with-db flag. After scaffolding, add azure-functions-db[postgres] (or another supported engine) to requirements.txt and wire up your bindings under app/.

---

Choosing a Stack

Stack	Scaffold Command	Packages	Use Case
API	afs new my-api	4	REST APIs, webhooks
DB API	afs new my-api + pip install azure-functions-db[<engine>]	5	CRUD with database
AI Agent	afs ai agent my-agent	3	LangGraph agents
Full	afs new my-api + pip install azure-functions-db[<engine>]	5	Production services

Start with the smallest stack that covers your needs. Add packages incrementally — the toolkit is designed for zero coupling between packages.