Creating Projects
Project Setup
Click the + New Project button from the Projects Dashboard to open the creation wizard. The first screen configures the basics of your project before you choose a tech stack.
Project Name
Enter a name for your project. The name is validated in real time — it must be a valid folder name (no special characters like <>:"/\\|?*) and must not conflict with an existing folder at the chosen location.
Project Location
The default directory where your project folder will be created. Click Browse to choose a different directory. The project is created as a subfolder using the project name.
Link to GitHub
Optionally select a GitHub account to automatically create a remote repository alongside your local project. Choose "None (Don't use GitHub by default)" to skip this.
Choose Platform
Select one of four platform tiles to determine the type of project you are building. Each platform shows a different set of framework and tech stack options on the next screen.
Blank
CLI tools, scripts, libraries
Desktop
Native desktop apps
Web
Browser-based apps
Mobile
Device applications
Build Type
Choose how the project should be initialized:
Fresh Install
Scaffolds a new project using standard framework tooling. Kaiju Hub AI runs the appropriate CLI commands (e.g., npm create vite@latest, cargo init, npx create-tauri-app) and installs all dependencies.
AI-Generated
Describe what you want to build and let AI generate the entire project. You provide a text description and optional reference images, then AI writes all the source code, configuration files, and project structure.
After configuring the project setup, click Continue to Tech Stack to proceed.
Blank Projects
Blank projects are for CLI tools, scripts, standalone applications, and language-specific starters. Choose your language or runtime and Kaiju Hub AI will scaffold a minimal project structure.
Language / Runtime Options
Some languages include version selectors. For example, selecting Python shows version options (3.10, 3.11, 3.12, 3.13, 3.14, 3.15). The selected version determines which Python interpreter is used to create the virtual environment.
Desktop Projects
Desktop projects create native desktop applications. Choose a desktop framework, a frontend framework for the UI, and optionally a database. All frontend frameworks come pre-configured with TypeScript.
Desktop Framework
The native framework that provides the desktop window shell and system-level APIs.
Frontend Framework
The UI framework that runs inside the desktop window. All options are pre-configured with TypeScript.
Database
Optional database integration for local or cloud data storage.
Web Projects
Web projects create browser-based applications. Start by choosing between Frontend Only or Full-Stack to determine which framework options are available. All frameworks are TypeScript-enabled by default.
Frontend Only
Standalone UI frameworks for single-page apps and static sites. Includes a Custom Packages button to add your own npm packages.
Full-Stack
Complete web frameworks with server-side rendering and API routes built in.
Styling
Database
Mobile Projects
Mobile projects create device applications for iOS and Android. Select a mobile framework, an SDK version, and optionally a database.
Mobile Framework
Choose the framework for your mobile application. An SDK version selector appears for frameworks that support multiple SDK targets.
SDK Version
For frameworks like Expo, select the SDK version (e.g., 52, 53, 54, 55). This determines the base dependencies and API compatibility.
Database
Optional cloud or local database for your mobile app.
Fresh Install
When you choose Fresh Install as the build type, Kaiju Hub AI scaffolds your project using standard framework tooling. After selecting your tech stack, click Create Project and Kaiju Hub AI handles the rest.
What Happens
1. Detect Package Manager
Kaiju Hub AI automatically detects which package manager is installed on your system (npm, yarn, or pnpm).
2. Run Scaffold Command
Executes the appropriate CLI tool for your chosen stack. For example: npm create vite@latest for React, cargo init for Rust, npx create-tauri-app for Tauri.
3. Install Dependencies
Runs the package manager install command to download and set up all required dependencies.
4. Initialize Git
Sets up a local Git repository with an initial commit.
5. Create GitHub Repo
If you linked a GitHub account, automatically creates a remote repository and pushes the initial commit.
The project appears in your Projects Dashboard once creation is complete, ready to open and start coding.
AI-Generated
When you choose AI-Generated as the build type, you describe what you want built and the AI writes the entire initial codebase. After selecting your tech stack, click Create Project to open the AI generation screen.
AI Model Selection
At the top of the generation screen, select an AI provider and model. You can choose from any configured provider (OpenAI, Anthropic, Gemini, Grok, DeepSeek, Ollama) and pick the specific model to use for code generation.
Project Description
Write a detailed description of what you want built. The more specific you are, the better the results. Include main features, pages, UI patterns, and any integrations you need.
Good Examples:
- An e-commerce website with product listings and shopping cart
- A task management dashboard with kanban boards
- A social media platform with user profiles and feeds
- A CRM system with contact management and analytics
Use the Enhance with AI button to automatically expand your description with more detail. You can also middle-click the microphone icon to dictate your description using voice input.
Sidebar Tabs
The left sidebar provides additional context for the AI:
Images
Upload reference images, mockups, or screenshots to show the AI what you want the app to look like. Drag and drop, paste with Ctrl+V, or click to add. Requires a vision-capable model to be selected.
Components
Attach reusable components from your Block Library. The AI will incorporate these into the generated project.
Files
Include reference files that the AI should consider during generation (e.g., existing code, design specs, or data schemas).
Generate
Click Generate Project in the top right to start AI generation. The AI will analyze your description and reference materials, plan the architecture, generate the file structure, and write all source code and configuration files. Progress is tracked in the Global Task Manager in real time.