V3XILI AI is a private AI workspace for students, makers and developers. It combines chat, vision, image generation, smart commands like /humanice and /x10think, and Plugging for connected tools. Built to help you study, code, research and create faster from one clean interface.
No reviews yetBe the first to leave a review for V3XILI AI
Maker
📌
Hey Product Hunt
I’m Carlos, the maker of V3XILI AI.
I built V3XILI AI as a private AI workspace for students, makers and developers who want one place to study, code, research and create content.
The current version includes:
• AI chat for study, coding and research
• Smart commands like /humanice and /x10think
• Vision Studio for image understanding
• Image Studio for AI image generation
• Plugging to connect tools and workflows
• A clean interface designed for desktop and mobile
The goal is not to build another generic chatbot, but a practical AI workspace that helps you actually get work done.
I’d love feedback on:
1. Which feature feels most useful?
2. What would you improve in the interface?
3. What integrations should I add next?
Thanks for checking out V3XILI AI
Report
How does the /humanice command actually work under the hood, is it just a prompt wrapper or does it fine-tune tone at the model level?
Right now, /humanice works as an instruction layer rather than a model-level fine-tune.
Under the hood, it takes the user’s original text, analyzes the intent and tone, then applies a structured prompt pipeline focused on making the output sound more natural, clear and human while preserving the original meaning.
So it is not currently fine-tuning model weights. I chose this approach because it is faster to iterate, safer to control, and easier to adapt across different models. In the future, I may explore tone-specific adapters or fine-tuned models, but for now /humanice is designed as a controllable prompt-command system inside V3XILI.
How does the /humanice command actually work under the hood, is it just a prompt wrapper or does it fine-tune tone at the model level?
@zbeydeharp17579
Great question.
Right now, /humanice works as an instruction layer rather than a model-level fine-tune.
Under the hood, it takes the user’s original text, analyzes the intent and tone, then applies a structured prompt pipeline focused on making the output sound more natural, clear and human while preserving the original meaning.
So it is not currently fine-tuning model weights. I chose this approach because it is faster to iterate, safer to control, and easier to adapt across different models. In the future, I may explore tone-specific adapters or fine-tuned models, but for now /humanice is designed as a controllable prompt-command system inside V3XILI.