Shubhamsaboo/awesome-llm-apps

🎯 What is Awesome LLM Apps?

A curated, open‑source catalog of 100+ ready‑to‑run AI agents, agent‑skills, and Retrieval‑Augmented Generation (RAG) applications. Every entry is a self‑contained template that you can clone, install, and run with a single API key (or locally, for the models that support it). The repo ships under the permissive Apache‑2.0 licence, so you’re free to copy, modify, or even commercialise the code.

📦 How it’s organised

Category	What you’ll find	Typical use‑case
Agent Skills	Small plug‑ins that give a coding‑assistant new abilities (e.g., Project Graveyard, Scope‑Creep Detector, Self‑Improving Skills)	Extend a code‑generation LLM with domain‑specific logic without writing the integration yourself.
Starter AI Agents	Single‑file, API‑key‑only agents (travel planner, blog‑to‑podcast, meme generator, medical‑imaging analysis, etc.)	Quick experiments or prototypes that run on Streamlit, FastAPI, or a simple script.
Advanced AI Agents	Multi‑step, tool‑enabled agents with memory, routing, and UI components (home‑renovation, VC due‑diligence, research planner, etc.)	Near‑production pipelines that combine web‑search, vector stores, and external APIs.
Always‑on Agents	Background services that run on a schedule or event (Hacker News briefing, release‑radar)	Automation that pushes updates to Slack, email, or other channels.
Multi‑agent Teams	Collections of agents that collaborate (competitor intelligence, legal bench, recruitment)	Complex workflows where several LLMs specialize and then aggregate results.
Voice AI Agents	Speech‑in / speech‑out bots (customer‑support, insurance claim intake)	Voice‑first interfaces that hook into real‑time transcription and TTS services.
Generative UI / MCP Agents	Agents that render interactive dashboards, component generators, or drive a real browser via the Model‑Context‑Protocol (MCP)	Turn natural‑language prompts into live UI widgets or browser actions.
Autonomous Game‑Playing Agents	Mini‑games where agents write, run, and play code (chess, tic‑tac‑toe, 3D Pygame)	Demonstrations of reasoning‑to‑code loops.
RAG Tutorials	End‑to‑end retrieval pipelines (local Llama 3.2, Gemini, DeepSeek, hybrid search, vision‑RAG)	Build your own document‑oriented chatbot with citations and self‑correction.

🚀 Getting started in 30 seconds

# Install a starter travel agent (requires an LLM API key)
git clone https://github.com/Shubhamsaboo/awesome-llm-apps.git
cd awesome-llm-apps/starter_ai_agents/ai_travel_agent
pip install -r requirements.txt
streamlit run travel_agent.py

Or add a skill to any coding assistant in a single command:

npx skills add https://github.com/Shubhamsaboo/awesome-llm-apps/tree/main/agent_skills/project-graveyard

The repo updates weekly with fresh templates – you can subscribe via the linked Unwind AI newsletter.

🛠️ Technical highlights

Model‑agnostic – works with Claude, Gemini, GPT‑4o, DeepSeek, Llama 3.x, Qwen, and other open‑source models.
Security‑first CI – each skill passes a security + evaluation gate before being merged.
MCP support – several agents use the Model‑Context‑Protocol to talk to real browsers, GitHub, Notion, etc.
Self‑improving skills – a subset can automatically rewrite themselves using Gemini and an internal ADK loop.
Docker‑friendly – most templates include a Dockerfile or can be containerised with a single docker build.

📚 Who should use this?

Developers who want a plug‑and‑play LLM agent without building the scaffolding from scratch.
Product teams looking for reference implementations of multi‑agent workflows.
Researchers interested in RAG, tool‑use, or autonomous game‑playing agents.
Educators who need concrete examples for teaching LLM‑driven automation.

🔗 Quick links

Run‑one‑now – see the “Run one now” section above.
Full template index – agent_skills/, starter_ai_agents/, advanced_ai_agents/, rag_tutorials/.
External showcase – Openwork browser automation agent (linked under Always‑on Agents).
Trendshift badge – repository highlighted as #1 of the day.

Bottom line: Awesome LLM Apps is a living toolbox of production‑ready LLM agents and RAG pipelines, designed for developers who want to ship AI‑powered features fast and legally (Apache‑2.0).