I’m launching a new series centered on super simple (micro) tutorials using Langflow. The goal is to quickly equip you with practical knowledge, inspiring the community to design their own use cases for Langflow and custom components.
What is it about:
- Practical Uses: The series will try to offer ideas on different ways to use Langflow and LLM pipelines in general.
- Emphasis on Custom Components: A major part of these flows revolves around custom components. You’ll learn how to design and integrate them into your flows.
- Resources: Each article provides a link to download the respective project. Use it to explore, modify, and understand the flow in-depth.
I hope the series is helpful and encourages you to discover creative ways to dive into the LLM landscape.
The first flow we will cover is a basic website reading flow that uses the prompt template to open space for multiple sources. Have fun!
Multiple Sources
Main Features
- Prompt Template: The project revolves around the use of a prompt template. This template serves as a blueprint for generating LLM prompts and creates input variables to be introduced from other components.
Objective
For small amounts of text, the prompt template is enough to enable users to retrieve data from various sources without relying on a vector store. In this project, input data comes from WebBaseLoader components, but you can imagine different loader types as well, with templates serving a diversity of purposes.
