From the course: Structured Vibe Coding with AI Coding Agents
Unlock this course with a free trial
Join today to access over 25,200 courses taught by industry experts.
Model selection for optimal vibe coding
From the course: Structured Vibe Coding with AI Coding Agents
Model selection for optimal vibe coding
- [Narrator] Just like human developers, coding agents need the right tools to work. Let's think about our own workflow. First, we need some sort of a IDE or a CLI environment. We need access to a code repository like GitHub, and some internet connection for docs and internet searches, and a terminal to run some comments. And most likely, we should be using Git for version control, and some planning tools like Jira. So if we need the coding agent to mimic our developer lifecycle, and in a typical developer lifecycle, we go through some of the workflow like the following, we will plan the work, we have to do some research online, we have to write and test code, we need to commit changes to Git, and then push it to GitHub, Trigger CI/CD, and to deploy to production. We want our coding agents to follow a similar cycle. This means we need to give it access to these tools an environment. The interface that connects coding agents to tools is called Model Context Protocol, or MCP. So, MCP is…