LLM
The AI brain powering tools like Lovable and Cursor.
What is LLM?
LLM stands for Large Language Model. It's the AI that reads your prompts and generates code. Claude (by Anthropic), GPT-4 (by OpenAI), and Gemini (by Google) are all LLMs. Vibe coding tools use them as their engine.
You don't interact with LLMs directly — Lovable, Cursor, and Bolt wrap them in user-friendly interfaces. But understanding what an LLM is helps when choosing tools or troubleshooting.
In Vibe Coding
Claude powers Cursor and Lovable. GPT-4 powers some features in Bolt. When you switch models in Cursor's settings, you're choosing which LLM generates your code. Claude tends to write better code; GPT-4 is better at general knowledge.
Example
For example: You ask Cursor to build a payment form. Cursor sends your prompt to Claude (the LLM), Claude generates the React code, and Cursor puts it in your file. The LLM is the brain doing the actual thinking.
Why this matters
This matters because LLM appears repeatedly when building, deploying, debugging, or connecting services in vibe coding workflows.
When you'll hit this in practice
You will usually run into LLM when working on What is Vibe Coding?.