awesome-llm-apps README

The project's README from github.com/Shubhamsaboo/awesome-llm-apps. As one markdown file: README.md.

Awesome LLM Apps

100+ open-source AI agents, agent skills, and RAG apps. Hand-built, tested end-to-end, Apache-2.0.

Clone it, ship it, sell it - 100% free and open-source

Works with Claude, Gemini, GPT, DeepSeek, Llama, Qwen and other open-source models.

Step-by-step tutorials on Unwind AI · Quick start · Browse all templates

Featured on Trendshift as the #1 repository of the day
Project Graveyard: an agent that autopsies your dead side projects Project Graveyard Insurance Claim Live Agent Team: voice claims settled in real time Insurance Claim Live Agent Team AI Fraud Investigation Agent: public records, cross-examined AI Fraud Investigation Agent
Self-Improving Agent Skills: skills that rewrite themselves against evals Self-Improving Agent Skills AI Home Renovation Agent: photo in, photoreal redesign out AI Home Renovation Agent Always-on HN Briefing Agent: it reads Hacker News while you sleep Always-on HN Briefing Agent

🙏 Thanks to our sponsors

TinyFish Community Programs: join the Students and Ambassadors programs
TinyFish
Become a Sponsor
Become a Sponsor

🚀 Run one now

Give your coding agent a new skill in 10 seconds:

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

Then ask it: "why do I never finish my side projects?"

Or clone and run any agent in 30 seconds:

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

📬 New templates drop weekly. Get them in your inbox on Unwind AI.

📂 Browse all templates

🧩 Agent Skills

Give your coding agent new abilities. One command to install, plain English to use. Every skill ships real code and passes a security + eval CI gate. Works with Claude Code, Codex, Cursor, and other coding agents. Browse all skills →

🌱 Starter AI Agents

Single-file agents that run with just an API key - a great place to start.

🚀 Advanced AI Agents

Production-style agents with tools, memory, and multi-step reasoning.

🛰️ Always-on Agents

Background agents that run on schedules or events, monitor changing context, decide what needs attention, and proactively deliver updates, artifacts, or actions.

🤝 Multi-agent Teams

Multiple agents collaborating to accomplish complex, cross-domain tasks.

🗣️ Voice AI Agents

Speech-in, speech-out agents using real-time voice APIs.

🖼️ Generative UI and Agentic Frontends

Agents that render interactive UI components, not just text: forms, cards, charts, editable plans.

🎮 Autonomous Game-Playing Agents

Agents that play games end-to-end: reasoning, strategy, and action.

♾️ MCP AI Agents

Agents that connect to external tools and data via Model Context Protocol.

📀 RAG (Retrieval Augmented Generation)

Retrieval pipelines, from simple chains to agentic and multi-source.

🔎 AI Browser Tools

Small tools that bring AI into everyday browsing.

💾 LLM Apps with Memory

Agents and chatbots that remember conversations and user state across sessions.

💬 Chat with X

Turn any data source into a chat interface.

🎯 LLM Optimization Tools

Reduce token usage, context size, and API cost without losing quality.

🔧 LLM Fine-tuning

End-to-end fine-tuning recipes for open-source models.

🧑‍🏫 AI Agent Framework Crash Courses

Deep-dive tutorials on the major agent frameworks.


⭐ Star the repo to get notified when new templates drop.

Deutsch · Español · français · 日本語 · 한국어 · Português · Русский · 中文

Apache-2.0 · See LICENSE · Fork it, ship it, sell it.

awesome-llm-apps (Shubhamsaboo/awesome-llm-apps)

vector-database
Searching Database…

What is this?A collection of 100+ open-source AI agent, agent skill, and RAG app examples - hand-built, tested end-to-end, and released under Apache-2.0 so you can clone, run, or even sell them.
Is this one app or a collection of separate projects?It is an index of standalone templates - each agent, skill, or RAG app lives in its own folder and is cloned and run independently, grouped into categories like Starter Agents, Multi-agent Teams, and RAG apps.
How do I get one of these agents running?Clone the repo and run any agent in about 30 seconds following that example's own instructions, or for Agent Skills, install a skill with one command and invoke it in plain English.
Do I need an API key to run these?Most Starter agents need just a single API key to get going, though several advanced and RAG examples (like the local Llama + Qdrant setups) run fully offline with no key at all.
Does every example work the same way?No - each folder is its own project with its own requirements, so some need one API key, some need several providers configured, and some are fully local; you set up each example individually rather than one global config.
Which AI models and coding agents does it support?The agents work with Claude, Gemini, GPT, DeepSeek, Llama, Qwen and other open-source models, and the Agent Skills are built to work with Claude Code, Codex, Cursor, and other coding agents.
Can I use these agents commercially?Yes - everything is Apache-2.0 licensed, and the README explicitly says you can clone, ship, and sell what you build with it.
What are Agent Skills in this repo?Agent Skills are small, installable abilities - like Project Graveyard or Scope Creep Detector - that give coding agents such as Claude Code, Codex, or Cursor a new capability through a one-command install and plain-English usage, and each one passes a security and eval CI gate.
How is the collection organized?Templates are grouped into categories such as Starter Agents, Advanced Agents, Always-on Agents, Multi-agent Teams, Voice AI Agents, Generative UI Agents, Game-Playing Agents, MCP Agents, and RAG apps, so you can browse by the kind of agent you want.
Is this list still actively updated?Yes, new templates are added weekly and announced through the Unwind AI newsletter.
Can I run any of these fully offline without API keys?Some can - for example the Local RAG Agent runs on Llama 3.2 with Qdrant and the Llama 3.1 Local RAG example works fully offline, though most Starter agents still need one API key.
Size
217 MB
Stars
142.1K
Updated
CommitOct 17
Age
2y
Commits · 90 Days