DEV Community: gezilinll

Proteus: The AI-native editor for multimodal creation

gezilinll — Fri, 02 Jan 2026 04:19:42 +0000

I'm building Proteus, an open-source multimodal editor (think Figma meets Notion, but AI-native) where AI writes most of the code while I focus on architecture, technical decisions, and quality control.

Why this matters:

In 2025, tools like Cursor and Claude can write good enough code in 80% of scenarios. The question isn't "Can AI code?" but "What becomes valuable when AI can code?" I believe it's system design, technical decision-making, and end-to-end ownership—not just knowing APIs.

What makes this different:

AI-native from day one: Every architectural decision prioritizes AI-friendliness. This isn't AI bolted on later—it's designed for AI collaboration from the first line.
Fully transparent: All code, architecture decisions, and lessons learned are public. I'm documenting the entire journey in weekly technical articles.
Real editor, not a toy: Phase 1 is complete with a working demo. You can create shapes, text, images, transform them, copy/paste, undo/redo—all the core editor capabilities.
Learning resource: If you want to understand how editors work (scene graphs, rendering, interaction systems) or how to structure code for AI collaboration, this is a live case study.

Current status:

✅ Phase 1: Core editing (scene graph, rendering, interaction, tools)

🚧 Phase 2: Multimodal elements (video, audio, web embeds)

📋 Phase 3: AI Agent integration (natural language → editor actions)

📋 Phase 4: Real-time collaboration

Try it: Live Demo

Code: GitHub

Articles: Tech Blog (4 articles so far, covering architecture, rendering, interaction design)

The experiment: What happens when you stop reviewing AI's code and instead focus entirely on architecture, problem diagnosis, and guiding AI through testing and context-building? That's what I'm exploring here.

Would love feedback from the HN community—especially from those building complex frontend apps or thinking about AI-native development workflows.

Alternative Shorter Version (if character limit is an issue)

Title: Proteus: An AI-native multimodal editor where AI writes 80% of the code

Description:

Building an open-source editor (Figma + Notion, AI-native) where AI writes most code while I focus on architecture and decisions.

Why: In 2025, AI can code—so what becomes valuable? System design, technical decisions, and ownership.

What's different:

AI-native from day one (not bolted on)
Fully transparent (all code + articles public)
Real editor (Phase 1 complete, working demo)
Learning resource (how editors work, AI-native architecture)

Status: Phase 1 ✅ | Phase 2-4 🚧

Demo: https://proteus.gezilinll.com/

Code: https://github.com/gezilinll/Proteus

Articles: https://github.com/gezilinll/Proteus/tree/main/articles

Experimenting with: What happens when you stop reviewing AI code and focus on architecture + problem diagnosis?

I have implemented a GPU version of Pica which is high quailty image resizer

gezilinll — Wed, 05 Mar 2025 01:06:32 +0000

🔗 GitHub Repo: https://github.com/gezilinll/pica-gpu
🔗 Demo: https://pica-gpu.gezilinll.com

Recently, while using Pica in a project, I noticed that both CPU and memory usage were quite high. When processing multiple images, the application would sometimes freeze or even crash. After reviewing the source code, I have to say that Pica's current CPU-based implementation is already optimized to an extreme level.

From what I observed, Pica's main optimization lies in its use of advanced filters for image scaling instead of conventional methods like nearest-neighbor interpolation (which prioritizes performance over quality). These filters have higher computational complexity but produce better results. Since filtering mainly controls how target pixels sample from the original image, and each pixel can be processed independently, this logic is inherently well-suited for GPU execution.

I have ported all of Pica's filtering algorithms to WebGL, creating a GPU-based version of Pica. In theory, using Compute Shaders could further improve performance, but since it requires WebGPU, which has compatibility concerns, I have not implemented that yet.

On the algorithm side, apart from mks2013, other filters theoretically support additional sharpening parameters. However, since our project primarily uses mks2013, which already provides excellent results, I haven't introduced sharpening logic to other filters yet—this is something I may improve in the future.

Currently, the GPU version of Pica achieves the same anti-moiré effect and sharpness as the original Pica while improving performance by 2-10×, with greater speedup for larger images. Additionally, because the GPU implementation avoids creating extra buffers, memory usage is lower. CPU load is also significantly reduced, which should help prevent performance bottlenecks.

Since Pica is designed as a CPU-based, JavaScript implementation, modifying it directly via a PR might be challenging for me. Instead, I have created a separate project.

I’d love for the community to collaborate on this! Any feedback or contributions would be greatly appreciated. Lastly, huge thanks to Pica for making our project possible! 😊

I am developing a Text Input Component based on Skia and Canvas

gezilinll — Sun, 14 Jul 2024 02:25:12 +0000

visit github

TextMagic is the next generation text component. Unlike native input and textarea components, it supports richer text effects and typesetting capabilities. By controlling text layout autonomously, it ensures consistent text display across different platforms and browsers. TextMagic follows a modular design approach, offering both an integrated component(@text-magic) for seamless integration and standalone components for specific needs: @text-magic/input for text input and @text-magic/renderer for text typesetting and rendering.

If anyone shares an interest in text or related fields, I welcome discussion and collaboration. I'm also in the process of learning in this area and would appreciate more feedback and assistance.

I am developing a Text Input Component based on Skia and Canvas

gezilinll — Sat, 13 Jul 2024 13:49:33 +0000

visit github
TextMagic is the next generation text component. Unlike native input and textarea components, it supports richer text effects and typesetting capabilities. By controlling text layout autonomously, it ensures consistent text display across different platforms and browsers. TextMagic follows a modular design approach, offering both an integrated component(@text-magic) for seamless integration and standalone components for specific needs: @text-magic/input for text input and @text-magic/renderer for text typesetting and rendering.

If anyone shares an interest in text or related fields, I welcome discussion and collaboration. I'm also in the process of learning in this area and would appreciate more feedback and assistance.