This isn’t just a list. It’s your personal LLM engineer’s toolbox, the kind of curated, no-fluff collection you’d bookmark after spending hours digging through GitHub, Stack Overflow, and research papers.
Think of it as the ultimate cheat sheet for anyone building real AI applications: from fine-tuning models to deploying agents, from crafting perfect prompts to keeping everything safe, fast, and measurable.
You’ll find tools that actually work:
- Fine-tune LLMs faster and cheaper (Unsloth, PEFT, LitGPT)
- Build RAG systems that don’t hallucinate (FastGraph RAG, Chonkie, BeyondLLM)
- Orchestrate smart, role-playing agents (CrewAI, LangGraph, AutoGen)
- Turn messy documents into clean data (Docling, Llama Parse, PyMuPDF4LLM)
- Keep your AI honest and secure (Guardrails, Garak, JailbreakEval)
- Make your apps run fast and scale smooth (vLLM, LightLLM, LiteLLM)
It’s organized by real workflow stages, not buzzwords.
No “AI magic” nonsense. Just practical libraries, clear descriptions, and direct links. Whether you’re training a model, scraping data, or monitoring a live agent, this repo has the right tool in the right place.
And here’s the kicker: it’s all open-source, free, and community-driven. No paywalls. No black boxes. Just powerful, transparent tools that let you build with confidence.
🧰 This is what happens when engineers stop chasing hype and start building what actually works.
If you’re serious about LLMs, whether you’re a researcher, developer, or builder, this is your go-to guide.Not because it’s flashy. But because it’s real.
