@sayginsamanVixentio: a visual workspace, from a 16-year-old founder in Miami
Hey everyone! I'm Saygin, a 16-year-old founder and developer based in Miami, Florida. I'm currently building Vixentio, a visual workspace designed to change how people build, collaborate, and work with AI.
I wanted to introduce myself, share what we're working on, and hopefully connect with other founders and developers in this community!
The problem that started everything
I've been building software with tools like Claude Code, Codex, and Cursor for a while now, and one frustration kept coming up.
AI has become incredibly good at writing code, but communicating exactly what you want it to build is still surprisingly difficult.
You can spend 20 minutes explaining a design change, describing where something should go, or trying to communicate what looks wrong with a UI. Sometimes the AI understands immediately. Other times, you end up going back and forth for hours.
I kept thinking: Why am I explaining something in paragraphs when I could just show it?
That question eventually became Vixentio.
What is Vixentio?
Vixentio is an AI-powered visual workspace that brings together visual thinking, project organization, team collaboration, and AI-assisted development.
The idea is simple: Show what you mean. Let AI build it.
Instead of constantly switching between screenshots, design tools, notes, chats, and coding agents, we're bringing these workflows together into one connected environment.
Here's what we're building:
- Visual AI workflows: Capture screenshots, annotate interfaces, highlight problems, and communicate changes visually to AI coding agents like Claude Code, Codex, and Cursor.
- Infinite Boards: Brainstorm ideas, sketch interfaces, create diagrams, and organize complex thoughts on a flexible visual canvas.
- Projects & Workspaces: Keep your ideas, research, visual references, discussions, and development context organized around what you're building.
- Workbench: A dedicated environment for bringing your visual ideas and development workflows together.
- Team Collaboration: Shared workspaces, conversations, and collaborative workflows designed to help teams build and communicate without losing context.
- AI-Powered Productivity: We're exploring how AI can understand not just what you type, but also what you see, draw, organize, and create.
The bigger vision
I don't want Vixentio to become just another note-taking app, another whiteboard, or another AI wrapper.
I believe the way we interact with software is fundamentally changing.
We're moving from manually doing everything ourselves to directing increasingly capable AI agents. But the interfaces we use to communicate with those agents haven't caught up yet.
Most AI interactions still happen through text boxes.
I think the next generation of AI tools needs to be more visual, contextual, and collaborative.
That's the future we're trying to build with Vixentio.
We're still early, constantly improving the product, experimenting with new ideas, and learning from the people using it.
I'd love to hear from this community!
For those of you using AI coding agents regularly:
What's your biggest frustration when trying to communicate design changes, UI improvements, or complex project requirements to AI?
And do you think visual interfaces will become a major part of how humans interact with AI agents?
I'd genuinely love to hear your thoughts, feedback, and experiences.
If you're building in the AI agent, developer tools, or productivity space, I'd also love to connect!
Check out Vixentio: vixentio.ai
— Saygin
Founder & CEO, Vixentio
Comments5
Welcome to Lounge, Saygin. The frustration of spending 20 minutes explaining a design change and still not getting it right is something a lot of people can relate to. Who feels that pain most today, solo builders or teams, and where are your first users coming from?
Thanks for the warm welcome! I think both solo builders and teams experience this frustration, but we're especially focused on small teams and design teams where communicating visual ideas clearly can make a huge difference. We're still in the early stages and don't have active users yet, but our current goal is to collaborate with designers, developers, and small teams to gather real feedback and improve Vixentio based on their needs. We're planning to bring in our first users through these partnerships and grow organically from there. Excited to be part of Lounge and connect with everyone here!
Welcome Saygin, impressive work at 16. Most of our product code is written with Claude Code, so here's my honest pain: the hardest part isn't describing the change, it's describing the state. "This button looks wrong" works only when the agent also knows which page, which viewport, logged in or not, light or dark theme, and what the data looked like. A screenshot annotated with that context (URL, viewport, the DOM element under the arrow) would save me more round trips than any canvas. The second pain is regressions: the fix lands, something two screens away breaks, and nobody notices for a day. So yes, visual will matter, but I'd bet the winning tools pair the picture with the machine readable context behind it. What do your current users mostly send through it, bugs or new designs?
Thank you, I really appreciate that! You actually described one of the biggest problems we're trying to solve with Vixentio. A screenshot alone isn't enough if the AI doesn't understand the context behind it, and that's exactly where we want to focus. Our vision is to make it easier to communicate not just what needs to change, but where, why, and under what conditions, so AI coding agents can understand the full picture with fewer back-and-forth iterations. We're still in the early stages and don't have active users yet, so we're currently looking to collaborate with developers and small teams to validate these problems and shape the product around real workflows. Initially, we're focusing more on UI changes and visual bugs, but regression detection is definitely an interesting direction to explore as well. Feedback like yours is exactly what we're looking for at this stage, and I think combining visual communication with deeper technical context could make a real difference in how developers work with AI.
5 comments
Welcome to Lounge, Saygin. The frustration of spending 20 minutes explaining a design change and still not getting it right is something a lot of people can relate to. Who feels that pain most today, solo builders or teams, and where are your first users coming from?
Thanks for the warm welcome! I think both solo builders and teams experience this frustration, but we're especially focused on small teams and design teams where communicating visual ideas clearly can make a huge difference. We're still in the early stages and don't have active users yet, but our current goal is to collaborate with designers, developers, and small teams to gather real feedback and improve Vixentio based on their needs. We're planning to bring in our first users through these partnerships and grow organically from there. Excited to be part of Lounge and connect with everyone here!
Welcome Saygin, impressive work at 16. Most of our product code is written with Claude Code, so here's my honest pain: the hardest part isn't describing the change, it's describing the state. "This button looks wrong" works only when the agent also knows which page, which viewport, logged in or not, light or dark theme, and what the data looked like. A screenshot annotated with that context (URL, viewport, the DOM element under the arrow) would save me more round trips than any canvas. The second pain is regressions: the fix lands, something two screens away breaks, and nobody notices for a day. So yes, visual will matter, but I'd bet the winning tools pair the picture with the machine readable context behind it. What do your current users mostly send through it, bugs or new designs?
Thank you, I really appreciate that! You actually described one of the biggest problems we're trying to solve with Vixentio. A screenshot alone isn't enough if the AI doesn't understand the context behind it, and that's exactly where we want to focus. Our vision is to make it easier to communicate not just what needs to change, but where, why, and under what conditions, so AI coding agents can understand the full picture with fewer back-and-forth iterations. We're still in the early stages and don't have active users yet, so we're currently looking to collaborate with developers and small teams to validate these problems and shape the product around real workflows. Initially, we're focusing more on UI changes and visual bugs, but regression detection is definitely an interesting direction to explore as well. Feedback like yours is exactly what we're looking for at this stage, and I think combining visual communication with deeper technical context could make a real difference in how developers work with AI.
Makes sense to start with UI changes and visual bugs, that is where the round trips hurt most. One idea for finding early users without a big launch: record 3 or 4 real before/after sessions (the vague prompt that failed vs the annotated screenshot that worked first try) and post them where Claude Code and Cursor users hang out. A side by side like that sells the problem better than any feature list, and the people who reply are your design partners. Good luck with it.