codesign

The description states that codesign is a visual autocomplete tool for interface design, operating within a canvas-based environment. It uses AI to suggest what the next step in a design could be, based on existing elements, layer names, and spatial relationships. Suggestions are staged separately from the original design, allowing designers to accept or reject them.

It supports three levels of fidelity:

• Base: Raw rectangles, text, layout
• AI Draft: Low-to-mid-fidelity interpretation
• AI HiFi: More detailed and polished direction

The tool is built using SvelteKit, with a canvas editor, scene graph, component registry, and a "design candidate" staging system. It integrates with Codex and GPT-5.6 Luna to operate on the canvas through an agent harness.

Inference: The product is described as a prototype, not a commercial offering. There is no evidence of revenue, pricing, or customer adoption.

The author positions codesign as a hybrid between traditional UI design and prompt-driven AI tools, inspired by tab completion in code editors. It aims to reduce the friction of AI-generated interfaces by allowing designers to remain on the canvas and maintain control over their design process.

Key claims:

• "What could visual autocomplete look like for interface designers?"
• "Designers continue working on a familiar canvas while Codesign sequentially suggests what their design could become next."
• "The AI basically does a 'visual autocomplete' based on existing elements in the working frame, layer names, and the placement of the elements in relation to each other."

Inference: The positioning is experimental and conceptual. It does not indicate any commercial traction or market validation.

The description states that codesign is intended for interface designers, particularly those who are familiar with canvas-based design tools and want to integrate AI into their workflow without losing agency or control.

Inference: The target customer is likely a subset of UX/UI designers, possibly in creative or product teams. No evidence indicates whether this group has been validated through market research, interviews, or usage data.

There is no evidence of a business model or pricing structure in the description. The project is described as a prototype built by one person over a week, with no indication of monetization or revenue streams.

The tool is built using:

• Codex
• GPT-5.6 Sol and Luna
• SvelteKit
• A canvas editor, scene graph, component registry, and staging system

It uses a "design candidate" staging system to separate AI suggestions from the original design.

Inference: The technical approach is experimental and relies heavily on AI agents and Codex integration. No evidence of scalability, performance, or production-grade delivery.

The project is described as a prototype built in one week, with no evidence of:

• Revenue
• Customers
• User adoption
• Product-market fit
• Prior versions or iterations

Inference: The product is at an early stage, likely experimental or academic in nature. There is no indication of traction beyond the author’s own use.

The description does not mention any competitors or direct market context. It references AI design tools that "start with a prompt and generate an entire interface at once", but does not name specific products or companies.

Inference: No competitive landscape is described, nor is there evidence of how this tool would differentiate from existing AI design tools in the market.

• Prototype-only: The product is described as a one-week prototype with no evidence of further development or traction.
• No commercialization path: There is no indication of a business model, pricing, or monetization strategy.
• Limited technical maturity: The author notes issues with visual bugs and generation quality, and that the tool is not yet production-ready.
• Self-reported only: All claims are from the author’s own account, with no independent verification.

1. What specific user feedback or use cases drove the development of this prototype?
2. Are there any early adopters or users beyond yourself?
3. How do you plan to transition from a prototype to a scalable product?
4. What are your thoughts on integrating with existing design tools (e.g., Figma, Sketch)?
5. Have you considered how this tool would be monetized or priced?