Shubhamsaboo/awesome-llm-apps

📚 What is Awesome LLM Apps?

Awesome LLM Apps is a curated, open‑source catalogue of more than 100 ready‑to‑run AI agents, agent‑skills, and Retrieval‑Augmented Generation (RAG) pipelines. Every entry is a complete, tested template that you can clone, install, and run with a single API key. The repo is organized into sections (Agent Skills, Starter Agents, Advanced Agents, Always‑on agents, Multi‑agent teams, Voice agents, Generative UI, Game‑playing agents, MCP agents, RAG tutorials, etc.) and each template ships with:

Real code (Python, Streamlit, or browser‑automation scripts) under an Apache‑2.0 licence – you can copy, modify, sell, or embed it in your product.
Model‑agnostic prompts that work with Claude, Gemini, GPT‑4/4o, DeepSeek, Llama, Qwen and other open‑source models.
A quick‑start guide – most can be launched in under a minute (e.g., pip install -r requirements.txt && streamlit run travel_agent.py).
Security & evaluation CI – each skill passes automated checks for safety and correctness before being listed.

The project is maintained by Unwind AI (the banner and tutorial links point to https://theunwindai.com) and is sponsored by companies such as Vorflux.

🚀 How to use it

Add a skill to a coding agent in seconds:

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

Then ask the agent a natural‑language question.

Run a full agent:

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

Most agents start with a single requirements.txt and a short Python script.

Subscribe for weekly updates – new templates are added regularly via the Unwind AI newsletter.

📂 What’s inside?

Category	Example templates (click to view)	What it does
Agent Skills	Project Graveyard, Scope Creep Detector, Self‑Improving Agent Skills	Small plug‑ins that give a coding‑assistant new capabilities (e.g., analysing abandoned repos, detecting scope creep, auto‑optimising themselves).
Starter AI Agents	AI Travel Agent, AI Blog‑to‑Podcast, AI Data Analysis	Single‑file agents that take an API key and run locally; great for learning or quick prototypes.
Advanced AI Agents	AI Home Renovation (photo‑in → render‑out), AI Fraud Investigation, AI Financial Coach	Multi‑step agents with memory, tool use, and external APIs; suitable for production‑style workloads.
Always‑on Agents	Hacker News Briefing, Release Radar	Background services that run on a schedule, monitor data sources, and push summaries to Slack/email.
Multi‑agent Teams	AI Competitor Intelligence Team, AI Legal Team, AI Services Agency (CrewAI)	Collections of specialized agents that coordinate via routing logic to solve complex, cross‑domain tasks.
Voice AI Agents	Insurance Claim Live Agent, Voice RAG (OpenAI SDK)	Speech‑in / speech‑out agents that use real‑time voice APIs.
Generative UI & Frontends	Generative UI Starter, AI Dashboard Canvas, AI Shadcn Component Generator	Agents that produce interactive UI components (cards, charts, kanban boards) as part of the conversation.
Autonomous Game‑Playing Agents	AI Chess Agent, AI 3D Pygame Agent	Agents that reason, plan, and act within a game environment.
MCP AI Agents	Browser MCP Agent, GitHub MCP Agent, Notion MCP Agent	Agents that talk to external tools via the Model‑Context‑Protocol (MCP), enabling live browsing, repo analysis, etc.
RAG Tutorials	Agentic RAG with Embedding Gemma, Corrective RAG, Vision RAG	End‑to‑end retrieval pipelines (local or cloud) with step‑by‑step explanations; many are fully offline.

🛠️ Technical highlights

Model‑agnostic – prompts are written to work with a wide range of LLM providers.
Apache‑2.0 licence – you can freely commercialise any template.
Security‑gated CI – each skill passes automated safety and evaluation tests before being listed.
MCP support – several agents use the emerging Model‑Context‑Protocol to call external services in a uniform way.
Streamlit & CLI interfaces – most agents expose a simple web UI (via Streamlit) or a command‑line entry point.

📦 Getting started quickly

# Clone the repo
git clone https://github.com/Shubhamsaboo/awesome-llm-apps.git
cd awesome-llm-apps

# Example: run the travel agent (requires an OpenAI/Claude/Gemini key in $OPENAI_API_KEY etc.)
cd starter_ai_agents/ai_travel_agent
pip install -r requirements.txt
streamlit run travel_agent.py

Replace the folder with any other template you’re interested in – the README in each subdirectory provides any extra setup steps.

🎯 Who should use this?

Developers who want a jump‑start for building LLM‑powered products.
Product teams looking for reference implementations of specific use‑cases (e.g., voice claim intake, RAG over PDFs, multi‑agent research).
Researchers / educators who need concrete, runnable examples of agentic reasoning, tool use, and retrieval pipelines.
Start‑ups that need a legally clear, Apache‑2.0‑licensed codebase they can ship or monetize.

📚 Further resources

Step‑by‑step tutorials – hosted on the Unwind AI site (https://theunwindai.com).
Trendshift badge – the repo is highlighted as a top‑trending AI project.
Sponsor links – Vorflux and a generic sponsor page are displayed in the README.

All information above is taken directly from the repository’s README; no external assumptions have been added.