Claude Bridge for Blender: chat on your own subscription and MCP for Claude Code

⚡ The gist in 5 seconds - The JackLuciano/claude-bridge-blender project has been published on GitHub — a bridge between a running Blender instance and Claude. - Two modes are claimed: chat directly inside Blender using your Claude subscription, and token-lean MCP operations for Claude Code. - System requirements, the Blender version, and installation steps are not disclosed in the description — check the repository. ### 🔍 What was found A repository called JackLuciano/claude-bridge-blender has appeared in the open, with the author's wording: "Drive a live Blender with Claude: in-Blender chat on your subscription + token-lean MCP ops for Claude Code." In other words, the project works with an already running, "live" Blender instance rather than exported files: the chat with the model is built directly into the editor, while Claude Code gets separate MCP operations optimized for token consumption. The mechanics are built on MCP — Model Context Protocol, the standard for connecting external tools to AI clients. Here, Blender acts as such a tool: Claude Code accesses it through the set of operations described in the project, while the built-in chat, according to the author, runs on the user's Claude subscription rather than separate API billing. The specific list of commands, configs, and the installation path are not given in the original description — they remain in the repository itself. ### 💡 Why it matters Over the past year, MCP has become the de facto standard for hooking AI assistants up to external software, and bridges to "heavy" desktop applications — such as Blender — have followed. For a 3D artist, this means an attempt to talk to the assistant in the language of the scene itself, without jumping into a browser or exporting intermediate files. A separate benefit is token savings: the more compact the operation descriptions, the cheaper and longer a long session in Claude Code lives. For now, this is a solo open-source project with no announced timeline for reaching stable channels, so it should be judged by its code and README rather than by the announcement.