Blender Agent Bridge is a free, open-source Blender extension that connects Blender to AI clients like Codex, Claude, and Cursor through MCP. It lets agents inspect scenes, gather visual evidence, and assist with workflows while keeping edits, script trust, paid generation, and approvals under the artist’s control. It Has integration with Meshy.ai, Tripo3D, Polyhaven and more.

I built Blender Agent Bridge after seeing a frustrating gap in AI creative tools: they could generate something impressive-looking, but when an artist needed to make a precise change, there often wasn’t an editable Blender scene, model, rig, or workflow underneath. The goal is not to replace Blender knowledge or remove the artist from the loop. It is to let AI assistants work inside a real Blender process: inspect the scene, gather viewport/render/playblast evidence, use structured tools, and help with longer workflows while Blender keeps control of edits, script trust, paid generation approvals, files, and credentials. I’m calling this a public beta because I want feedback from real Blender users, technical artists, and people building MCP-based workflows. I’d especially love to hear where the workflow feels useful, where it feels clunky, and what safety or control expectations you’d want before using this on real projects.

Open-source tools that make AI and Blender work together more seamlessly are exciting—congrats on the launch! 🎉

Been using BlenderMCP for the Claude side of this. How does the bridge compare if I already have Cursor in the loop, same kind of scene control or more of a one-way export?

Open-source tools that make AI and Blender work together more seamlessly are exciting—congrats on the launch! 🎉

Been using BlenderMCP for the Claude side of this. How does the bridge compare if I already have Cursor in the loop, same kind of scene control or more of a one-way export?

Blender Agent Bridge

Open-source MCP bridge for Blender AI workflows

Open-source MCP bridge for Blender AI workflows