Building on the previous episode, we look at refactoring our background job into a more maintainable object and provide context to the LLM so that we can chain together responses for a more conversational experience.

In this episode, we explore the practical use of Campfire Bots for webhook applications. We'll start by setting up an AI Image Generator bot that interacts with our Ruby on Rails application, and then connect it to a self-hosted Large Language Model (LLM). Our initial goal is to get the system working. After that, we'll focus on refactoring the code to improve its safety and maintainability.

In this episode, we'll have a look at a way where we can use fixtures to manage the data in our development environment, tests and a demo environment. We'll dive into some complications and nuances with handling fixtures in ActionText and ActiveStorage.

Setting up a Rails Engine with the TailwindCSS and Hotwire via Importmaps can be rather difficult. There's a lot of undocumented steps to take and it can be overwhelming to approach these problems. In this episode, we look at a step by step approach to both adding TailwindCSS and Turbo & Stimulus into a Rails Engine.

In this episode, we look at setting up our Campfire instance to be deployable with Kamal instead of the Once CLI. This has a few benefits over the standard deployment mechanism if you are going to be tweaking and making your own changes to the Campfire code. We'll also look at taking in the incoming changes from Once to rebase into our Kamal setup.