
Skills, now in Framer. Start designing on the canvas and let the Agent infer your layout and design rules, iterating as it generates. Skills are dynamically invoked, too, so the Agent decides which ones to read based on your prompt. Canvas × Skills hits different. Give it a go!
Agents in @framer now support all 3D Transforms. This lets you iterate on all sorts of creative effects at an incredible pace, while all generated 3D layers remain directly selectable and editable on the Canvas. framer.com/updates/agent-3d-…
I’m in the final stages of building the next version of @framer’s agent harness. It’s getting so good that I’ve become convinced that any teams paying attention will reverse engineer and copy its key innovations within a month. Not giving any hints, but you’ll be able to tell if agents suddenly get faster, more responsive, and use less tokens. More soon.
The current branch name and context indicator are now neatly displayed below the input in the @framer agent, making it a bit easier to see where you’re working and cleaning up the input.
Framer Agent’s model picker now includes Reasoning and Fast Mode, so you can control how much a model thinks before building and, when supported, how fast it works.
We’re bringing Video Shaders to @framer, starting with a new version of our Chromatic Aberration shader, which now supports images, videos, and mouse interactions, all-in-one. Add glass-like curvature, dispersion, and animations to your assets with just a few clicks. Web-ready.
benjamin retweeted
Framer AI agent is climbing the OpenRouter ranks very fast only 7 weeks after launch.
Rank: 7
Tokens: 472B
openrouter.ai/rankings#apps
GPT 5.6 Terra and Luna are 50% off through August 14.
Plus 32 Agent updates and 14 Canvas updates.
All in the July update. Now in Framer.
framer.com/updates/july-upda…















