mrbizarro/Phosphene — Run MiniMax Hailuo H3 and LTX-2.5 video generation locally on a Mac.

Description and use

Phosphene enables creators to run MiniMax Hailuo H3 and LTX-2.5 video generation models locally on Macs using MLX, eliminating the need for CUDA, cloud services, or API keys while supporting joint audio-video generation and character LoRA training. Its practical adoption is streamlined through a one-click Pinokio installation, allowing users to integrate these capabilities into their content production pipelines with minimal setup friction. With strong recent activity, 183 GitHub stars, and a high production readiness score, the project is well-suited for serious pilot implementations, though final review of its license and security posture is recommended.

Problem

mrbizarro/Phosphene helps creators research, generate, repurpose, and publish content with less manual work.

Outcome

High for an OSS candidate: recent activity, adoption, and ecosystem signals are strong enough for a serious pilot.

Workflow

• research and produce content
• repurpose assets across channels
• automate a creator production pipeline

Evidence

Run MiniMax Hailuo H3 and LTX-2.5 video generation locally on a Mac. Joint audio+video, character LoRA training, one-click Pinokio install. MLX — no CUDA, no cloud, no API key.

Source