Phoenix: A Claude-Powered Text-to-3D Pipeline
Double_Ad_6033 · reddit · 2026-07-20
The open-source project Phoenix chains the text-to-3D workflow into a seamless pipeline: it uses ComfyUI + Trellis for image-to-3D, and then leverages Claude to assemble the results in Blender, ultimately creating a pipeline where text goes in and a usable mesh comes out.
The author highlights a clear division of labor:
- ComfyUI / Trellis handles generation
- Blender handles assembly
- Claude manages operation and guidance
They also acknowledge current limitations: thin and complex structures like grass and branches tend to break, so those parts are kept procedural rather than forced through the model. It works best for more compact subjects like tree trunks, statues, or deer. The author is actively seeking community feedback on missing capabilities.
More from coding & agent
- Agent Field Kit launches $29 security runbooks for production AI agents — alvelda · 2026-10-27
- Cloudflare Durable Objects validated as top architecture for AI agent WebSocket servers — tristanbob · 2026-10-27
- GitHub Next's Chopin prototype explores real-time multiplayer agentic planning — max__drake · 2026-10-27
- Veynod: An Open-Source Shared Workspace Where AI Agents Collaborate Like a Team — TKOWNS · 2026-10-27
- TMLR paper: enforced thinking doesn't reliably make LLM agents disclose key info to users — SharonYixuanLi · 2026-10-27
- LLM-written Rust library bbtree verified: speed claims hold, memory savings reverse — lmoroney · 2026-10-27