Drafts
Drafts
Drafts are AI-generated content waiting for your review. When you ask Giant Context to create something—whether it's a page, a form, an email, or a knowledge base article—the result doesn't go live immediately. Instead, it lands in Drafts first. This acts as a staging area between the AI building your content and it officially becoming part of your project.
The drafts list
Navigating to Drafts from the left sidebar gives you a complete overview of all your AI-generated content. When you ask Giant Context to build something—whether it's a new website page, a registration form, an email, or a knowledge base article—the result doesn't go live immediately. It lands in your Drafts table first, giving you full control to review, refine, or reject the output before publishing.
Tip. You don't need to wait on this screen while your draft is generating. You can safely navigate away and come back later when your content is ready for review.
Accepting and Rejecting Drafts
Rejecting Drafts. When you reject a draft, the actual pages, forms, or other content the AI created are cleanly removed so they don't clutter your project. The draft record itself stays in your list marked as "Rejected" so you always have a history of what was generated.
Methods for Creating Drafts
Giant Context integrates AI at every layer of the platform, allowing you to generate content rather than starting from a blank page. Whether you are creating a single page or a connected set of forms and emails, all AI-generated content lands in Drafts for your review before going live.
There are three primary ways to initiate a draft:
- Manual generation: Describe what you want in plain language using the chat interface.
- Programmatic generation: Use the API or SDK (
POST /drafts/generate-new) to trigger drafts via integrations and automation. - Automatic generation: The Mind engine continuously analyzes your project and surfaces ideas for new content, which you can kick off with a single click.
Console Chat
Track Generation Progress
When you ask Giant Context to create something, the AI begins drafting immediately. Monitor the status in real-time—from kickoff to Ready—or navigate away and come back later to review the results.
Programmatic Generation (API & MCP)
Mind Autonomous Engine
Rejected drafts are free. If a draft isn't exactly what you're looking for, reject it and try again with a new prompt. There's no cost to rejecting—the generated content is cleanly removed, giving you the freedom to iterate until the result is perfect.
Automated Drafts from Mind Ideas
Giant Context includes a continuous intelligence layer called Mind. Rather than waiting for manual prompts, Mind actively analyzes your existing project content to identify gaps and surface strategic opportunities for new material.
When you approve a proposed idea from Mind, it automatically triggers the draft generation process. The AI immediately begins building the requested content, placing the finished result directly into your Drafts queue for final review and publishing.
Training the Mind Engine. Mind ideas include priority levels, such as 'Low' for gap detection. Dismissing irrelevant ideas isn't just about clearing your queue—it actively trains the engine to deliver better, more accurate proposals in the future.
FAQ
The logic dialog
Every draft maintains a complete history of its generation process. You can review the chat log to see the original request and track progress messages from when the draft was kicked off to when it was ready. If you decide not to use the generated content, the draft record stays in your list marked as "Rejected" so you always have a history of what was proposed.
Understand How Your Content Was Built
Get complete transparency into the generation process. The draft logic dialog reveals your original prompt, an overall quality score, and the exact sequence of tool calls the AI used to assemble your pages, forms, or emails.
No, draft generation happens in the background. You can navigate away and come back when it's ready.
Consider using plain 'Accept' first to fine-tune the content in the editor and check mobile responsiveness before going live.
Use specific prompts with rich detail. Instead of 'Create a pricing page', specify the exact tiers, prices, and sections you want.
Yes, combine content types by asking for related content together, like 'a signup page with a registration form', to generate a connected multi-item draft.