AI Agents with Flowise

Explore credentials used in this course and create service credentials for each lecture. Access free tiers where possible, and prepare for services like OpenAI that may require a budget.

set up OpenAI credentials by creating an account on platform.openai.com, top up billing, generate a secret key, and connect it to Flowise to run OpenAI API models in your flows.

Set up Google search API with a programmable search engine, note the search engine ID, and create cloud project to enable the custom search API and generate an API key.

Sign up for serp api to perform google searches on the free tier, then copy your api key and add it as a serp api credential in flow credentials.

Create an Upstash account and database, then copy the vector endpoint and configure credentials in Flowise using the vector API token.

Learn to set up Pinecone, a vector API, create an account and API key, add credentials in Flowise, and build a configured index with embedding model and host for upserts.

Create a Supabase account, form an organization, and set up a database on the free plan, then use the direct connection info to connect the manager to the system.

Explore agent flows by examining start, agent, and human-in-the-loop nodes, then manage variables, tools, and retrieval augmented generation with proprietary knowledge bases in Flowise through practical projects.

Create a simple chatbot using an LLM in Flowise, integrate it into a web page with a template and script, and tune prompts and configuration for website use.

Master the start node and flow initiation with chat or form input, manage flow state variables and ephemeral memory, and configure the lm node with system, user, and developer messages.

Explore the agent node and direct reply node, compare with the LM node, and learn to use tools, knowledge bases, vector stores and embeddings, document stores, and Rag.

Build a blog writing team with a blog writer using web search tools (Google Custom Search, GPT-4) and a blog evaluator that scores the article from 1 to 10.

Explore retrieval augmented generation by attaching a knowledge base to an agent via document loaders, text splitters, embeddings, and a vector store, with a record manager to keep information current.

build a rag chatbot using OpenAI embeddings and a vector store with a Supabase record manager, loading winter sports PDFs and chunking them for efficient retrieval.

Learn to use built-in tools like web search, calculator, and date/time in Flowise, and build custom tools such as a text metrics analyzer for article writer and analyzer workflows.

Understand how variables in agent flows use double curly bracket notation and dot notation to access flow data. See how a company name variable drives a scheduling flow for appointments.

Explore how human in the loop adds control over tools and workflows by requiring confirmation for tool use and workflow progress, with hotel and blog examples.

Explore how to build your own agent flows and agents and integrate them into your projects, and provide feedback to help shape future AI courses.