Describe your idea → AI builds a complete, runnable web application
Describe your web application in detail — what it does, who it's for, key features, design preferences, etc. The AI will generate a complete, runnable webapp with HTML, CSS, and JavaScript.
Select which AI models to use for generating your webapp. Models are loaded from your configured API providers.
Primary model for code generation and architecture decisions.
Secondary model for quick tasks like parsing and formatting.
Model used for image generation or vision tasks (required by the pipeline).
Takes your webapp idea and generates a complete project with HTML, CSS, and JavaScript. The AI agent plans the architecture and implements all files.
Run the full build pipeline. The AI will read your idea and generate a complete, runnable web application.
Already built your webapp? Write notes describing what you'd like to change, fix, or improve. The AI will read your notes and update the existing project accordingly.
Run the update operation to apply your notes to the existing project. The AI will read your feedback and modify the generated code accordingly.
Reads your update notes and applies changes to the existing webapp project files.
Instead of "make it better", say "increase the font size of headings to 1.5rem"
Explain what happens vs. what should happen: "Clicking delete removes all tasks, should only remove the selected one"
List the most important changes first so critical fixes are addressed even if the AI runs out of context
You can run multiple update cycles — refine a little at a time for best results
Run the pipeline to generate your webapp, then it will appear here.
Run the pipeline to generate project files.
Run the pipeline to generate the project README.
Download your generated webapp files as a ZIP archive. You can download the entire project or specific subdirectories.
Or download a specific path:
View the current state of your project's Git repository. Initialize a repo to start tracking changes.
Stage all changes and create a new commit. This automatically runs git add -A before committing.
View, create, and switch between branches to manage different versions of your webapp.
Switch or create a branch:
View the recent commit history for the current branch.
Common workflows combining multiple Git operations.
Initialize repo (if needed), then commit all current changes with a timestamped message.
Commit current work, create a new "experiment" branch, and switch to it.
Commit all changes, then download the entire project as a ZIP archive.
View token usage and estimated costs for all AI model calls made during this session. Usage is tracked per model and includes prompt tokens, completion tokens, and cost estimates.
No usage data yet. Run a pipeline operation to generate usage data.
Usage breakdown for individual task sessions. Each pipeline run or update creates a separate task session with its own usage tracking.
Task-level usage will appear here after pipeline runs.
Prompt tokens are the input sent to the AI model. Completion tokens are the output generated. Both contribute to cost.
Costs are estimated based on the model's pricing. Actual billing may vary slightly depending on your API provider.
Usage is tracked per session. Each browser session gets its own usage counter that resets when a new session starts.
Click on any task session link to view detailed usage for that specific AI operation in real time.