No reviews yetBe the first to leave a review for VRMA-lab
Maker
📌
Hi Product Hunt! I’m the solo maker behind VRMA Lab 👋
I built VRMA Lab after repeatedly running into a simple problem: making a small correction to a character animation often means opening a complex 3D application and learning a workflow much larger than the edit itself.
VRMA Lab is my attempt to make those small edits feel immediate. It runs entirely in the browser, with no account, upload, or backend required. Models, motions, project data, and exports stay on the user’s device.
It provides three workflows:
• Human Workbench — visually adjust bones, expressions, and gaze with real-time preview and A/B comparison.
• Quick Converter — import a humanoid FBX animation and export a validated VRMA without entering the editor.
• Agent Mode — let AI agents inspect, edit, review, validate, and export motion through structured commands instead of learning to click a graphical interface.
Agent Mode is the most experimental part. Rather than asking an AI to navigate dozens of buttons through repeated screenshots, VRMA Lab gives it structured state and a fixed four-view visual review surface. The feature already works, but it is still a new direction, and I would genuinely love help testing it.
To be honest, reaching the right users has been more difficult than building the product itself. The initial launch has received very limited visibility, which also means I have not yet collected enough real-world usage feedback. I am still trying to understand whether the main barrier is the way I explain the product, the onboarding experience, or simply reaching such a specialized audience. If you try VRMA Lab, even telling me where you felt confused, stopped using it, or expected something different would be incredibly valuable.
VRMA Lab is free, local-first, and open source under Apache-2.0.
I’d especially love feedback from VRM creators, animation-tool developers, and AI agent builders:
What is the first character-motion task you would try, and where does the workflow still feel confusing?
Live demo: https://yuban834.github.io/vrma-...
GitHub: https://github.com/YuBan834/vrma...