

Projekt is a visual workspace designed specifically for design engineers that provides a unified environment for working with AI coding agents. It wraps AI coding agents in a real-time preview, an editable DOM, and a prompt-first interface, allowing users to write naturally, see results instantly, and refine visually.
The platform enables users to run multiple AI coding agents side-by-side with live preview, inline editing, and file management capabilities. Users can connect their preferred AI agents including Claude Code, Codex, Gemini CLI, or Opencode, bringing their own API keys without vendor lock-in. The workspace provides a clean, minimal interface with file explorer, live preview, and chat functionality integrated into a single window.
Projekt works by allowing users to connect their AI coding agents using their existing API keys, then open any project folder without setup or configuration. Users describe what they want through prompts and see live previews update in real-time as their agents write code. They can select any element, edit inline, and tweak styles visually before committing and deploying their work.
The product enables builders to go from idea to shipped product without waiting on anyone, providing tools for prompt-based development with instant visual feedback. It supports working on any project with full control over the development stack and AI models used.
Projekt is built specifically for design engineers and people who know that design is the differentiator in software development. It assumes users are the whole team and need to work efficiently from concept to deployment using AI coding assistants.
admin
Projekt is specifically crafted for design engineers and people who recognize that design is the key differentiator in software development. It targets builders who go from idea to shipped product without waiting on anyone, assuming users are the entire team. The tool serves individuals working on web development projects who want to leverage AI coding agents with visual feedback and prompt-based workflows.