New Claude Connector Update Is INSANE!
Anthropic released nine new official connectors for Claude desktop, with the headline addition being an official Blender connector built by the Blender team itself. The connector allows Claude to read and modify 3D scenes directly through natural language, eliminating the need for manual MCP server setup. Similar connectors for Adobe tools like Photoshop and Illustrator are also rolling out.
Summary
Anthropic released nine new official connectors for Claude desktop, marking what the presenter calls the biggest creative tool drop the company has ever done. The headline addition is an official Blender connector built by the actual Blender development team — not a third-party GitHub project — making it a first-class, supported integration shipped directly inside Claude desktop.
The Blender connector gives Claude two core capabilities: reading a 3D scene and writing to it. On the reading side, Claude can inspect every object, material, light, camera, and node in a scene. On the writing side, it can add or delete objects, fix materials, reorganize folders, and run cleanup tasks — all triggered by plain-language instructions from the user. Practical examples include renaming hundreds of poorly labeled objects automatically, identifying high-polygon-count assets that are slowing down a scene, deleting unused materials, and explaining complex node trees in plain English step by step.
A key distinction the presenter emphasizes is how much easier setup has become. Previously, connecting Claude to Blender required manually configuring an MCP (Model Context Protocol) server, downloading files from GitHub, and running terminal commands — a process most non-technical users abandoned. Now the process is reduced to enabling the connector in Claude desktop settings, installing the official MCP add-on in Blender, and clicking 'Connect to Claude' in Blender's side panel.
The presenter is also careful to set realistic expectations: Claude will not replace skilled 3D artists and still struggles with complex generative tasks like building a human face from scratch or handling mesh topology correctly. Its strength lies in automating repetitive, tedious tasks — the 'do this 100 times' work — while the human remains in a directing and reviewing role.
Beyond Blender, the presenter highlights that the same connector pattern is being extended to Adobe's creative suite, including Photoshop, Illustrator, Premiere, and After Effects, as well as tools like Figma. The broader argument made is that the gap between having an idea and producing a finished creative asset is shrinking rapidly, and those who adopt these tools now will have a significant advantage over those who wait.
Key Insights
- The presenter argues that the new official Blender connector eliminates the previously painful MCP server setup process, reducing what used to require GitHub downloads and terminal commands down to two clicks inside Claude desktop and one add-on installation in Blender.
- The presenter claims Claude can now explain complex Blender node trees — which can involve hundreds of interconnected boxes — step by step in plain English, describing it as 'like having a Blender teacher sitting next to you,' something that previously took years of learning to interpret.
- The presenter explicitly states that Claude is not going to replace a 3D artist, noting it still gets scale wrong and mishandles topology, and frames its value specifically as automating boring, repetitive tasks rather than generative creative work.
- The presenter describes a practical scene-optimization use case where Claude scans a slow-running kitchen scene, identifies a fridge with 2 million polygons, and recommends reducing it to 50,000 — work that would otherwise require the user to manually dig through menus and count data.
- The presenter draws a direct historical parallel to Claude Code adoption, arguing that people who learned Claude connectors early will 'run circles around everyone else' in six months, just as those who jumped on Claude Code six months ago are doing now.
Topics
Full transcript available for MurmurCast members
Sign Up to Access