RAG Agents: Build Apps & GPTs with APIs/MCP, LangChain & n8n

Learn the basics of LMS, vector databases, embeddings, chunks, and top-k results, plus API basics, ChatGPT fundamentals, and building your own GPT with practical tools.

Explore how large language models work from pre-training to reinforcement learning, compare open-source and closed-source options like Llama and ChatGPT, and cover tokens, token limits, and DirectX retrieval augmented generation.

Learn how large language models use function calling to talk to tools via APIs, vector databases, and embeddings, enabling browsing, calculations, a Python interpreter, and image or video generation.

Learn retrieval augmented generation by using embedding models and vector databases to store and search documents, manage chunking and top-k results, and enable in-context retrieval with LLMs.

Learn the basics of ChatGPT and the OpenAI playground, covering GPT-5, test time compute, instant thinking options, templates, the agent builder, and audio and images via the OpenAI agent SDK.

Master ChatGPT basics: interface, models, settings, and GPTs, plus the OpenAI Playground. Explore canvas, vision, web search, deep research, and custom GPTs to build apps with APIs.

Explore test time compute, a thinking-through approach that lets models like chatgpt generate chain-of-thought for math, code, and science questions, balancing accuracy with cost and speed.

Learn to build your first rag app hands-on in ChatGPT, prep data from pdfs, html, and csv, and train GPTs on your writing style to generate leads.

Transform static html pages into a rag chatbot by scraping a single url to markdown and crawling links to build an llm-trained knowledge base.

Train ChatGPT to mimic your writing style using retrieval-augmented generation, by uploading sample texts, building a vector database, and guiding posts across blogs and social media.

Prepare data from PDFs by converting to markdown and storing essential information in a vector database; craft a short system prompt for efficient ChatGPT use with open-source, private workflows.

Learn to build a local rack application with Ollama and Anything LLM, configure chunking and model setup, and explore local agent skills like web search and charts.

Install anything locally, connect it to ollama, and run RAC apps with AI agents using the anything LM interface and Lans db locally.

Learn to set chunk size and chunk overlap for embeddings to optimize document embedding in a vector database, balancing accuracy, token limits, and pricing.

Explore practical agent capabilities in anything LM, including memory, document summarization, web search, scraping, chart generation, and MCPs, with flexible prompts and local workflow tools.

Configure Ollama locally by selecting VRAM‑matched models with tool calling, and manage them via GitHub commands. Build a local rec app using chunking, prompts, and a vector database.

Discover the OpenAI playground to build rag apps with text-to-speech, audio models, vector stores, and the response api, paying only for output tokens.

Install node.js to run flowwise locally and test llm apps. Explore hosting options from flowwise free plan to self-hosted deployments on AWS, Render, or Docker.

Resolve local node and flow wise installation issues by managing node versions with nvm for Windows, downgrading to compatible 18–20 releases and switching versions as needed.

Install flow wise locally with node.js and the command prompt, then use npm to install and start flow wise, launching the local server on port 3000 and update as needed.

Explore the flow wise interface, learn to navigate chat flows, agent flows, marketplace, tools, credentials, and document stores. Build agents from scratch using nodes, vector stores, and custom tools.

Build a RAG chatflow that scrapes web content, creates embeddings, and queries an in-memory vector store with LangChain's conversational retrieval QA chain, HTML splitter, and Brave Search API document loader.

Export and import a jet flow as json to share and reuse workflows, load a downloaded flow into a new canvas, and customize or export your own for distribution.

Explore the Flowise tool-agent to enable a dual agent that connects any API and LLM, using Open Router, Claude, and Compose IO with memory and diverse tools.

Learn to build a tool agent with RAG by integrating Pinecone as a vector store, embeddings, and a retriever to upsert and query document chunks across Postgres or Supabase.

Learn how to create a Pinecone vector database, troubleshoot embedding options, and apply a practical workaround to create an index with text embeddings and custom configurations.

Install Node.js locally, manage versions with npm, and set up n8n to automate node-to-node data flows. Create workflows with Google Drive inputs, Pinecone storage, and web scraping legally.

Install and run n8n locally by setting up node.js, managing versions with nvm, and exploring self-hosting with Docker and cloud options; try free for two weeks.

Master node version management with nvm to fix n8n installation errors, ensuring admin rights on your local machine and selecting compatible node versions using nvm list, install, and use.

Discover n8n basics by building workflows with triggers and actions, exploring nodes, models, and chat-based ai agents, and wiring data through json inputs, memory, and tools.

Explore building ai powered workflows in n8n with the cloud add-on update, templates, and data tables, to create a knowledge assistant using API keys and a vector database.

Build a rag bot for lead generation with n8n, Pinecone and Google Sheets, using Google Drive triggers to upsert data into Pinecone and capture leads.

Convert a website into a rag chatbot by scraping pages with HTTP requests, converting HTML to markdown, and indexing in pinecone with OpenAI embeddings, while honoring robots.txt and sitemaps.

Wrap up the n8n section by using addons, triggers, and actions to build a rack application that inserts data from Google Drive, and review legal web scraping and workflows.

Learn to self-host and deploy RAG-Agents across devices, build rack applications, and explore selling them, with Replit, WordPress embeds, and branding.

Host flow wise apps on render to deploy cloud chat flows from a forked GitHub repo, configure environment variables and persistent disk, and choose a free or starter plan.

Build a rag chatbot for a client using Flowwise tool agent, document stores, and pinecone embeddings; train it with medical data and a German system prompt.

embed a chatbot into a wordpress site by using the vp code plugin and an html snippet with the chatbot script placed in the footer for all devices.

Explore advanced flow settings for a LangChain / LangGraph app in Flowise, including view messages, feedback, leads capture, rate limits, starter prompts, speech-to-text, and Google Sheets integration.

Explore practical self-hosting of n8n using render, hostinger, and other options; compare free trials, pricing, and docker setups, with persistent disks and remote triggers.

Build and publish a standalone n8n rag lead bot as a hosted web app with a public URL and optional authentication. Share or embed the chat for client access.

Learn to integrate Nadan chatbots into web pages via embed chat, WordPress plugin, and CSS branding, with testing on Replit and step-by-step customization of code and styles.

Learn to sell rag agents by identifying the right customers, prototyping tailored chatbots, and delivering live demonstrations with return on investment calculations, offers, guarantees, and smart pricing.

Explore hosting and self-hosting RAG agents with render or flowwise, test for free, and deploy branded chatbots—embed in WordPress, collect leads, and even sell standalone apps.

Learn how the model context protocol (MCP) standardizes LLM access to APIs via a USB-C style host-client-server model, enabling easier function calling and dynamic self discovery.

Connect ChatGPT with n8n via webhook to trigger AI agent workflows, using Google Sheets, Calendar, Gmail tools, and production testing to automate leads, emails, and data flows.

Explore cache augmented generation (CAG) and its contrast with retrieval-augmented generation (RAG), weighing fast latency and simple design against accuracy limits and volatile, short-lived caches in production.

GraphRAG uses knowledge graphs to connect entities across vector chunks for more accurate, context-aware results; increase top_k or chunk size for cheaper prices and more relevant outputs.

LightRAG provides a fast, cost-effective alternative to GraphRAG with cheaper API calls and training, offering a Python installation for quick setup and comparable information retrieval.

Learn contextual retrieval, a chunking strategy from Anthropic that improves RAG accuracy by contextualizing chunks, using embeddings, a vector database, prompt caching, and fast free models like Grok and Gemini.

Choose the right rag strategy based on data size and accuracy, starting with standard rag in a vector database, then consider contextual retrieval, cag, or graph rag and light rag.

Master the model context protocol (mcb), a function calling–style bridge linking an lm to APIs and their functions. Explore prompts, webhooks, prompt caching, and contextual retrieval for efficient cross-app workflows.

Explore best practices and risks in building AI agent apps, including OpenAI API keys, data privacy, jailbreaking, prompt injections, copyright, costs, and European regulatory considerations.

Understand the privacy risks of using Telegram as a trigger in n8n and learn to secure workflows with an if node that restricts access by chat ID and username.

Examine jailbreaking LLMs and AI agents via prompts, including many-shot and zero-shot attacks, base64 and hashing tricks, and image prompts with noise patterns, with cross-model examples and safety implications.

Examine data poisoning and backdoor attacks on LLMs across pre-training, instruction training, and fine-tuning, and discuss data security, privacy, and function calling with tools like Midjourney, Dall-E, and Adobe Firefly.

Navigate copyrights and intellectual property for AI-generated data with agents, including OpenAI safeguards, llama licenses, and rules for training data usage, selling, and distributing outputs.

Explore the EU AI Act and GDPR compliance for chatbots, including risk-based classifications, data privacy, transparency, consent, and EU data residency with OpenAI to ensure responsible, compliant AI apps.

Understand the trade-offs between API-based and local LLMs, balancing costs with reliability, while applying data privacy, copyright, security, and compliance practices in building RAG apps.

Explore how this course guides you to build rack applications and RAG workflows with vector databases, prompt engineering, and multi-agent frameworks.