fastapi-rag 0.1.0

Enterprise AI backend template generator for FastAPI and RAG systems

fastapi-rag

fastapi-rag is an installable Python CLI that generates production-grade FastAPI backends for AI applications, RAG systems, internal copilots, and enterprise SaaS platforms.

It is intended to behave like a real developer tool, not a local scaffold script.

What It Does

After installation:

fastapi-rag new myapp
cd myapp
docker compose up --build

The generated project includes:

• FastAPI application structure
• JWT authentication
• SQLAlchemy async setup
• provider abstractions for DB, cache, queue, LLM, and vector stores
• RAG ingestion and query pipeline
• Redis and Celery integration
• Docker Compose startup
• health endpoints, metrics, and observability wiring
• tests and developer docs

Repository Structure

src/fastapi_rag/
├── cli/
│   └── main.py
├── generators/
│   └── project_generator.py
├── renderer/
│   └── jinja_renderer.py
├── templates/
│   └── full_rag/
└── utils/
    └── package_resources.py

Installation

Local development install

pip install -e .

Package build

python -m build

This generates:

• wheel in dist/*.whl
• source distribution in dist/*.tar.gz

Install after publishing

pip install fastapi-rag

Quick Start

Generate a new backend:

fastapi-rag new myapp

The CLI will ask for:

1. project name
2. LLM provider
3. vector database
4. database provider
5. cache provider

Then it will:

1. load the packaged template
2. copy it into ./myapp
3. render all Jinja variables
4. output a complete backend project

Run the generated project:

cd myapp
docker compose up --build

Supported Provider Defaults

The packaged template currently supports or models:

• LLM: echo, openai, ollama
• Vector DB: qdrant, chroma, pgvector, pinecone
• Database: postgresql, mysql
• Cache: redis, dragonfly
• Queue: celery

Architecture Overview

The generated backend uses a modular architecture with:

• app/api/
• app/core/
• app/db/
• app/modules/
• app/providers/
• app/services/
• app/workers/

The generator itself is separated from the scaffold so packaging and runtime template access stay reliable after installation.

Package-Safe Template Loading

Templates are loaded with importlib.resources, not fragile relative paths.

That means fastapi-rag can work after:

• pip install -e .
• pip install fastapi-rag
• wheel installation
• source distribution installation

Development Workflow

Install local dependencies:

pip install -e .[dev]

Run tests:

pytest

Build distributions:

python -m build

Validate package metadata:

twine check dist/*

CI and Release Automation

This repository includes GitHub Actions workflows for:

• CI test and build verification
• version and release automation
• TestPyPI publishing
• PyPI publishing

Workflows:

• .github/workflows/ci.yml
• .github/workflows/release-please.yml
• .github/workflows/publish-testpypi.yml
• .github/workflows/publish-pypi.yml

Publishing

Build locally

python -m build

Upload to TestPyPI

twine upload --repository testpypi dist/*

Install from TestPyPI

pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple fastapi-rag

Versioning

This repository is configured for conventional commits and release-please.

Expected examples:

• feat: add anthropic provider prompt option
• fix: include templates in sdist build
• docs: improve publishing guide

Release automation updates:

• version in pyproject.toml
• release PR
• git tag
• GitHub release
• changelog

Documentation

• docs/ARCHITECTURE.md
• docs/USER_GUIDE.md
• docs/FIRST_FEATURE.md
• docs/PUBLISHING.md
• CONTRIBUTING.md

License

This project is released under the MIT License.