Member-only story
I’ll be honest, for the longest time I used Claude the same way I use Google — type something in the box, get an answer, move on. Never really stopped to think about what was actually happening underneath. Then a coworker mentioned we were considering the API for a new feature, and I realized I couldn’t answer a pretty basic question: what actually is Claude?
So here’s me working that out, mostly for my own benefit.
Claude is Anthropic’s family of language models. Anthropic is the company — founded in 2021, with an unusually heavy focus on AI safety baked into how they build things. “Claude” itself isn’t one model, though. It’s more of a product line: several models at different sizes and price points, all from the same lineage, tuned to do different jobs well.
The thing that took me a minute to actually click: no matter how you’re talking to Claude — through the chat window, through your own code calling the API, through an agentic tool running semi-autonomously — you’re talking to one of these same underlying models. The wrapper changes. The engine doesn’t.
Once that clicked, some of my earlier confusion made more sense. I kept mixing up “Claude” the model with the various things you access it through. Claude.ai is the chat interface, the one most…