Metadata-Version: 2.4
Name: kickstart-ai
Version: 0.1.2
Summary: KickstartAI: a lightweight scaffolding engine that turns natural language prompts into full project structures.
Author-email: Christopher Cruz <cchris2004@gmail.com>
License: MIT
Project-URL: Homepage, https://github.com/cruz209/kickstart-ai
Project-URL: Issues, https://github.com/cruz209/kickstart-ai/issues
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: torch>=2.0.0
Requires-Dist: transformers>=4.35.0
Requires-Dist: huggingface_hub>=0.20.0
Requires-Dist: requests>=2.28.0
Requires-Dist: PyYAML>=6.0
Requires-Dist: openai>=1.0.0
Dynamic: license-file

🚀 kickstart-ai

AI-powered code scaffolding that actually respects your brain.
kickstart-ai turns a short natural-language request into a clean, debuggable project skeleton — Flask apps, FastAPI services, Streamlit dashboards, RAG demos, whatever. It generates the folder structure, starter code, configs, templates, and even validates the output so you don’t get hallucinated junk.

Not a Cursor clone. Not a “press button → production SaaS” hallucination machine.
Just a reliable launchpad that gets you from idea → working scaffold fast, while keeping you in control of the real coding.

✨ What kickstart-ai Does

Parses vague user intent

Builds a meta-prompt describing the ideal project

Uses pluggable LLM backends (OpenAI first, HF fallback)

Generates a file-tree JSON

Writes the whole project to disk

Validates imports, directories, and missing files

Warns you about errors before you even touch the code

🔧 Why It’s Useful

kickstart-ai gives you the code you would've written anyway, just faster:

Perfect for prototypes, demos, hackathon starters

Great for students who need structure but still want to code

Ideal for devs who hate boilerplate

Avoids hallucinated garbage — you get clean TODOs, not fake logic

🧠 Philosophy

kickstart-ai is a launchpad, not a replacement for developers.
It handles the boring scaffolding so you can focus on the real engineering.

🗺️ Roadmap

v0 — core scaffolding engine (done)

v1 — template packs + stronger validator

v2 — intent classifier + safety layer

v3 — plug-and-play local models (llama.cpp, GGUF)

v4 — interactive generation / iterative refinement
