Hey HN,
I've built an open-source CLI tool that generates production-ready full-stack templates for AI/LLM applications. It's designed to cut down on boilerplate so you can focus on building features – think chatbots, assistants, or ML-powered SaaS.
Install via pip install fastapi-fullstack, then run fastapi-fullstack new for an interactive wizard to customize your project.
Key features:
Backend: FastAPI with Pydantic v2, async APIs, auth (JWT/OAuth/API keys), databases (async PostgreSQL/MongoDB/SQLite with Alembic), background tasks (Celery/Taskiq/ARQ), rate limiting, webhooks, Redis caching
Frontend: Optional Next.js 15 (App Router, React 19, Tailwind, dark mode, i18n) with real-time WebSocket chat UI
AI/LLM: Integrated PydanticAI for type-safe agents, tool calling, streaming responses, and conversation persistence
Observability: Logfire for tracing everything from API requests to agent runs; plus Sentry/Prometheus
CLI: Django-style management commands with auto-discovery (e.g., user creation, DB seeding, custom scripts)
DevOps: Docker Compose, CI/CD templates (GitHub Actions/GitLab), Kubernetes manifests
20+ configurable integrations to pick what you need – no bloat
Repo: https://github.com/vstorm-co/full-stack-fastapi-nextjs-llm-t...
Inspired by tiangolo's FastAPI templates and others, but with a stronger AI focus, modern frontend, and more enterprise-grade options out of the box.
Screenshots, demo GIFs, architecture diagrams, and full docs in the README.
This has sped up my own projects a lot – curious about yours: What pain points do you hit with full-stack AI setups? Any features to add (e.g., more LLM frameworks like LangChain coming soon)? Contributions welcome!
Thanks
Backend: FastAPI with Pydantic v2, async APIs, auth (JWT/OAuth/API keys), databases (async PostgreSQL/MongoDB/SQLite with Alembic), background tasks (Celery/Taskiq/ARQ), rate limiting, webhooks, Redis caching Frontend: Optional Next.js 15 (App Router, React 19, Tailwind, dark mode, i18n) with real-time WebSocket chat UI AI/LLM: Integrated PydanticAI for type-safe agents, tool calling, streaming responses, and conversation persistence Observability: Logfire for tracing everything from API requests to agent runs; plus Sentry/Prometheus CLI: Django-style management commands with auto-discovery (e.g., user creation, DB seeding, custom scripts) DevOps: Docker Compose, CI/CD templates (GitHub Actions/GitLab), Kubernetes manifests 20+ configurable integrations to pick what you need – no bloat
Repo: https://github.com/vstorm-co/full-stack-fastapi-nextjs-llm-t... Inspired by tiangolo's FastAPI templates and others, but with a stronger AI focus, modern frontend, and more enterprise-grade options out of the box. Screenshots, demo GIFs, architecture diagrams, and full docs in the README. This has sped up my own projects a lot – curious about yours: What pain points do you hit with full-stack AI setups? Any features to add (e.g., more LLM frameworks like LangChain coming soon)? Contributions welcome! Thanks