Awesome LLM Apps: Over 100 Open-Source AI Projects to Learn From

There is a repository that collects more than 100 open-source AI projects -- from simple agents to multi-agent teams, voice agents, MCP integrations, and RAG applications -- all in one place, with code you can open, read, and run.

What it is

Awesome LLM Apps is a curated collection of over 100 open-source AI agent, agent skill, and RAG projects, released under the Apache-2.0 license. The repository is not a framework or a unified platform. It is a shelf of independent example projects, each in its own folder with its own code, dependencies, and setup instructions. You find a problem that matches yours, open the code, see how the pieces connect, and adapt it.

The examples are organized into categories: starter AI agents, advanced AI agents, multi-agent teams, voice AI agents, MCP agents, RAG tutorials, agent skills, and generative UI agents. The repository includes examples using Claude, Gemini, GPT, DeepSeek, Llama, Qwen, and other open-source models, so you can find something close to whatever stack you are already using. At the time of research, the repository showed approximately 128.6 thousand GitHub stars.

Why it matters

• Real problems, not just chatbots. The examples go beyond basic Q&A. They include a project that analyzes abandoned side projects, an AI fraud investigation agent that cross-references public records, and a live insurance claim agent team that handles intake by voice.
• Complete app structure. For newcomers, seeing a full AI application -- model connection, tool integration, data handling, and UI -- in one place is more valuable than reading documentation in isolation.
• Range of complexity. The repository spans from single-file starter agents to multi-agent orchestration, voice interfaces, MCP integration, and generative UI. You can start simple and work upward.
• Model flexibility. Because examples use multiple model providers, you are not locked into one ecosystem. You can find patterns that match the models you already have access to.

How it works

Running an example is straightforward. You clone the repository, navigate to the specific project folder you want to try, install its requirements, and launch it. Many projects use Streamlit for a quick web interface. For example, the travel agent quickstart involves cloning the repo, entering the starter_ai_agents/ai_travel_agent directory, running pip install -r requirements.txt, and launching with streamlit run travel_agent.py. Of course, you still need to add your own API keys and read the README for each specific project, since every example has its own configuration requirements.

Caveats

The project count, star count, and model list can change over time. The "hand-built, tested end-to-end" claim is the author's description, not an independent audit. Each project uses different dependencies, API keys, and external services. Code that runs is not the same as code that is production-ready. Before using any example in a real product, you need to evaluate security, API costs, data privacy, error handling, and operational stability. The repository is best treated as a learning and prototyping resource, not a template library for production deployment.