A curated collection of small LangChain/LLM demos and utilities. Each subproject is self‑contained with its own README.md or notebook and focuses on a single capability (chat, retrieval, summarization, SQL, PDFs, audio, code generation, etc.).

• Use per‑project README.md files for setup and run instructions
• Common prerequisites: Python 3.9+, virtual environments, API keys where required
• Typical run command for apps: streamlit run app.py inside the project folder

• blog generation/ — Streamlit app generating short blogs via local Llama 2 (ctransformers)
• calories health/ — Gemini Vision nutrition analyzer from food images; optional SQL demo
• chat multiple documents/ — Chat over multiple PDFs with FAISS + Gemini
• conversational Q&A chatbot/ — ChatOpenAI conversational demo with simple memory
• invoice data extractor/ — Extract structured JSON from invoices using LangChain parsers
• langchain audio to text/ — Whisper transcription with optional post‑processing via chat models
• langchain basic qa/ — RetrievalQA over a single PDF using in‑memory doc store
• langchain summarization/ — Web page summarization using stuff, map_reduce, refine chains
• LLM generic app/ — Notebook pipeline: PDFs → OpenAI embeddings → Pinecone → QA
• news research tool project/ — RockyBot: URLs → FAISS index → QA with sources
• project codebasics Q&A/ — CSV FAQs → FAISS + Instructor embeddings → PaLM Q&A
• Q&A chatbot using LLM/ — Minimal single‑turn Q&A with OpenAI completions
• sqldb tshirts/ — MySQL inventory Q&A via SQLDatabaseChain and few‑shot prompts
• text summarization/ — Notebook‑based summarization for text/PDFs
• wordpress code assistant/ — Generate WordPress PHP code with QA cross‑check

• The file project_structure.text enumerates the repository’s current layout and was used to compile this overview: langchain-projects\project_structure.text:1-110

• API keys: many apps require keys (OpenAI, Google Generative AI, etc.). See individual project READMEs
• OS specifics: some projects include Windows‑specific dependencies (e.g., python-magic-bin for unstructured)
• Notebooks: open with Jupyter/VS Code; run cells sequentially

Each app uses third‑party APIs and models subject to their terms. Ensure you have rights to any data you process and comply with applicable licenses and policies.