Running Qwen 3.8 27B Locally to Drive Blender via MCP and Build a 3D Llama
jacek2023 · reddit · 2026-09-06
A full local setup for driving Blender 5.x with Qwen 3.8 27B via MCP: install the Blender MCP extension, hook it up through pi with pi-mcp-adapter, and configure .mcp. to launch blender-mcp from the v1.0.0 repo via uvx on localhost:9876. The post shows the model's actual planning—it weighs procedural building vs sculpting vs metaballs for making a 3D llama, then procedurally assembles a stylized chibi llama from primitives. A reproducible workflow for running open local models against desktop creative software.
More from coding & agent
- Enough ThreeJS slop: using AI to reverse engineer a 15-year-old game's physics engine — JasonBotterill · 2026-09-06
- HelpMeOut launches an agent-to-agent Q&A network for business questions — AI_spell · 2026-09-06
- OpenAI developer says internal coding tool Astra pulled plans forward by six months — The Decoder · 2026-09-06
- A complete AGENT.md: one dev's tuned global ruleset for coding agents — alexcovo_eth · 2026-09-06
- Dev proposes open-source self-healing agent loop: trajectory eval turns failures into skills — EnchantedHawk · 2026-09-06
- Polygres turns your Postgres into a hybrid search context layer for AI agents — Scobleizer · 2026-09-06