Recipe Categories
RAG Applications
| Recipe | Description | Stack |
|---|---|---|
| RAG Chatbot | Chat with your documents | LangChain + ChromaDB |
| Multi-PDF QA | Ask questions across PDFs | LlamaIndex + Qdrant |
| Web RAG | Search and summarize websites | Crawl4AI + Claude |
| SQL RAG | Natural language to database | Text2SQL + PostgreSQL |
AI Agents
| Recipe | Description | Stack |
|---|---|---|
| Research Agent | Autonomous web research | CrewAI + Tavily |
| Code Review Agent | Automated PR review | Claude + GitHub API |
| Data Analysis Agent | Analyze CSVs with AI | PandasAI + GPT-4 |
| Customer Support Bot | Intent detection + routing | LangGraph + Claude |
Multi-Model Apps
| Recipe | Description | Stack |
|---|---|---|
| Model Router | Route to cheapest model | OpenRouter + FastAPI |
| A/B Testing | Compare model outputs | Promptfoo + 3 models |
| Ensemble | Combine multiple model answers | Voting + aggregation |
Specialized
| Recipe | Description | Stack |
|---|---|---|
| Voice Assistant | Speech-to-text + LLM + TTS | Whisper + Claude + ElevenLabs |
| Image Analyzer | Describe and analyze images | Claude Vision + Streamlit |
| Email Writer | Context-aware email drafts | Claude + Gmail API |
| Meeting Summarizer | Transcribe + summarize | Whisper + Claude |
Each Recipe Includes
recipe-name/
├── README.md # Setup instructions
├── requirements.txt # Dependencies
├── app.py # Main application
├── config.py # Configuration
├── .env.example # Required API keys
└── tests/ # Basic testsKey Stats
- 6,000+ GitHub stars
- 50+ complete recipes
- 10+ categories
- Full source code (not snippets)
- Active maintenance and new recipes weekly
FAQ
Q: What is Awesome LLM Apps? A: A collection of 50+ production-ready AI application examples with complete, runnable source code covering RAG, agents, multi-model apps, and specialized AI tools.
Q: Is Awesome LLM Apps free? A: Yes, all recipes are open-source under MIT license. Some recipes require paid API keys.
Q: Can I use these recipes in production? A: Yes, recipes are designed as production starting points. Add your own error handling, authentication, and scaling as needed.