fastai-stack

Build production-ready, AI-first FastAPI backends in minutes.

fastai-stack is an interactive CLI generator for modern Python teams that want strongly typed APIs, async-first architecture, and built-in paths for LLM apps, embeddings, background jobs, observability, and container deployment.

Why fastai-stack

• AI-native backend scaffolding, not generic boilerplate
• FastAPI + Pydantic v2 conventions out of the box
• Opinionated project structure for scale and maintainability
• Interactive and non-interactive modes for both humans and CI
• One command to scaffold full backend architecture

Core Features

• Typer-based CLI with clean UX and smart defaults
• Jinja2 template engine with conditional generation
• Modular API layout (app/api/v1/endpoints/*)
• Pydantic settings model for environment-driven config
• Async-ready backend structure
• Optional AI endpoints for chat and embeddings
• Optional vector DB integration (pgvector, weaviate)
• Optional task queue support (celery + redis)
• Optional monitoring support (sentry, prometheus)
• Docker + Docker Compose with CPU/GPU profiles
• Poetry-based dependency management
• Test + CI scaffolding included

Install

pip install fastai-stack

Quick Start

Interactive Mode

fastai-stack create myapp --interactive

Non-Interactive Mode

fastai-stack create myapp \
	--non-interactive \
	--db postgres \
	--auth jwt \
	--tasks celery+redis \
	--ai openai \
	--vector-db pgvector \
	--docker gpu \
	--monitoring prometheus \
	--frontend htmx+vite

CLI Options

fastai-stack create <project_name> [options]

• --interactive / --non-interactive
• --db: postgres | mongodb | sqlite
• --auth: none | jwt | oauth2
• --tasks: none | celery+redis
• --ai: none | openai | langchain | huggingface
• --vector-db: none | pgvector | weaviate
• --docker: cpu | gpu
• --monitoring: none | sentry | prometheus
• --frontend: none | htmx+vite
• --output-dir <path>

Generated Project Structure

<project_slug>/
|- app/
|  |- core/
|  |  |- config.py
|  |  |- deps.py
|  |  `- security.py
|  |- api/v1/endpoints/
|  |  |- health.py
|  |  `- ai/ (optional)
|  |- models/
|  |- schemas/
|  `- crud/
|- tests/
|- migrations/ (postgres)
|- docker-compose.yml
|- Dockerfile
|- .env.example
|- pyproject.toml
`- .github/workflows/ci.yml

AI Capabilities

• OpenAI chat endpoint template with SSE token streaming
• Embeddings endpoint starter template
• Hugging Face sentence-transformers friendly defaults
• LangChain-ready integration stubs

Dev Experience

• Consistent settings loading through pydantic-settings
• Clear separation of API, core, schemas, and crud layers
• Ready-to-run local development setup
• Starter tests for smoke validation

Local Development

poetry install
poetry run pytest
poetry run fastai-stack --help

Running a Generated App

Inside the generated project:

cp .env.example .env
poetry install
poetry run uvicorn app.main:app --reload

Open docs: http://127.0.0.1:8000/docs

Docker Usage

docker compose up --build

If you choose --docker gpu, the generated compose config includes NVIDIA device reservations.

Publishing This CLI

poetry build
poetry publish

Roadmap Direction

• Richer auth blueprints (fastapi-users full wiring)
• Production-ready DB model templates
• More AI provider templates and eval-ready scaffolds
• Full release automation for PyPI + GitHub Releases

License

MIT