Ngo Thuan (Copax)
About the author

Copax – Crafting AI, from Idea to Product

“Where AI takes shape”

1.5 Million+combined downloads and model runs across platforms (estimated total)
970K+runs on Tensor.art (handle: SDG)
310K+downloads on Civitai, 3,500+ followers
418K+uses on Mage.space
~69self-trained models on Civitai
153virtual characters, 93,800+ chats on Fizzz

Figures aggregated/estimated from the platforms (checked 10/2026), updated periodically.

Story

I am Ngô Thuần, known in the international AI community by the nickname Copax.

I started as a SharePoint developer, spending about 4 years building enterprise management solutions on the Microsoft cloud platform. When Stable Diffusion emerged, I left my corporate job to focus entirely on generative AI.

Since 2023, I have released self-trained models on Civitai, later expanding to Tensor.art (under the handle SDG), SeaArt, Mage.space, and Hugging Face. The Copax TimeLess and SDG-Realistic model series have been used hundreds of thousands of times. I then moved on to building AI apps from custom ComfyUI workflows, creating over 150 virtual characters on Fizzz, and returning to my core programming strengths: building backends, credit systems, multi-provider API gateways, and desktop software for AI music and video processing.

SDG (sydigi.com) is where I bring it all together: software, APIs, and knowledge, making AI practical and accessible for everyone.

Expertise

  • AI Image Model Training: Checkpoints and LoRAs for SD1.5, SDXL, Flux, Pony, Illustrious, Qwen, Z-Image.
  • Workflows & AI Apps: Designing ComfyUI workflows and packaging them into concept-based image generation apps.
  • API Infrastructure: API gateways for multi-provider key management, load balancing, error handling, multi-worker servers, and providing APIs for website and tool integration.
  • Turning 'Vibe Code' Apps into Products: Backend development, user management, credit systems, and private server deployment.
  • Desktop Software for AI Music & Video: Stem separation, mastering, hook finding, and beat-sync video editing.
  • Virtual Characters & AI Content: Images, music, and video.

Mission

Bringing AI from the lab to real users: user-friendly tools, stable APIs, fair pricing, and transparent knowledge sharing for both the Vietnamese and international communities.

Journey

  • ~4 years SharePoint developer, enterprise solutions on Microsoft Cloud
  • 2023 Self-trained models published on Civitai, Tensor.art, SeaArt, Mage.space, Hugging Face
  • Next AI apps built from ComfyUI workflows, 150+ virtual characters on Fizzz
  • Now Backends, API gateways, AI music & video desktop software – all in one place at SDG