When you ask Claude to explain something complex — compound interest, decision trees, chemical reactions — the answer used to be text. Maybe with a code artifact on the side. Now Claude builds the visualizations right into its response.
What’s new?
Anthropic rolled out a feature yesterday that builds on ‘Imagine with Claude,’ previewed last fall. Claude decides on its own when a visual would be helpful — or you can ask directly: ‘Draw this as a diagram’ or ‘Visualize how this changes over time.’
The visualizations appear inline, right in the middle of the response. No side panel, no separate window. And they’re interactive — you can change variables and see the results immediately.
How does it work technically?
Claude generates HTML and SVG behind the scenes. You don’t see any of that — it just looks like an embedded widget. The visualizations are temporary and evolve with the conversation. As the discussion develops, the graphics adapt or disappear.
Who is this for?
Pretty much everyone. A few examples from the announcement: interactive compound interest curves, clickable decision trees, and a periodic table you can explore. I find it especially exciting for education and data analysis — when you can not just read about connections but actually see and interact with them.
Availability
The feature is rolling out as a beta across all pricing tiers — including free users. It’s enabled by default.
My take
This is one of those moments where you realize: chat as an interface is far from done. Instead of just polishing text responses, the conversation itself becomes an interactive medium. OpenAI recently did something similar with interactive math visualizations in ChatGPT — but Claude’s approach is broader and not limited to specific subjects.
I’m curious how this feels in practice. For technical explanations especially, this could be a real game changer.
Sources: