Gemini Builds a Website in Under a Minute: Five New Features Tested

Google Gemini can now generate and publish a professional-looking website in about a minute, according to hands-on testing by Your AI Workflow. That headline capability leads a batch of five new features spanning Google Sites, PDF handling, Gmail, Docs, and Slides — each of which trims a workflow down to prompts, while leaving a few manual steps in place.

The website builder works from a canvas. Users upload images and logos, which Gemini hosts with unique links in webp format, then describe what they want — say, a storefront — and the tool generates the design. In testing, Gemini produced six product listing images for hoodies, t-shirts, and backpacks, in light and dark variants, each carrying a minimal logo. The result: a storefront with a "minimal premium design," concise copy, and smooth transitions between products.

Publishing happens through an export button that pushes the site directly to Google Sites. Sites can be made public or restricted to the user's organization. There's one structural catch: because Gemini publishes the code as an embed, the top navigation bar in Google Sites cannot be removed. That's the kind of detail that decides whether a quick-built storefront passes for a hand-built one.

What does the PDF feature actually fill?

The PDF-filling feature works inside the chat interface and pulls data from Google Workspace apps. In one test, filling out an invoice, Gemini retrieved the recipient's company name, address, itemized prices, total amount due, and a tax ID from Gmail threads, dropped the tax ID into the notes section, and even recorded the payment method as a wire transfer.

It stops short of finishing the job. Completed documents aren't attached to emails automatically; users must download the file and upload it to a Gmail draft themselves. A separate PDF filler tool was also mentioned, offering sharing options, adjustable access permissions, and open notifications.

How well does AI search work in Gmail?

"Ask Gmail" replaces the standard search bar and returns AI-generated overviews for specific questions. Asked about the status of an invoice, Gemini pulled from multiple email threads and notifications, summarizing billing details, scope, and current status — and even identified when the recipient had opened the shared invoice, using email notification data.

The limits are query-shaped. Ask "Who do I need to follow up with to complete invoice payments?" and you get no AI overview at all. The feature performs well for narrow, person-focused questions but struggles with broad or generic ones.

Can you draft documents by voice?

In Google Docs, the "create with Gemini" button on iPhone now supports voice input. Users dictate a prompt; Gemini generates a detailed document. In one test it returned five video concepts for a merch launch, complete with style suggestions, audio ideas, hooks, and detailed timelines. A source settings tab lets users specify which apps or web sources Gemini should draw on, keeping the output aligned with what they actually need.

What about automated slides?

Google Slides generation runs from a single prompt, with documents or existing presentations attached from Google Drive as style and content references. Before building, Gemini confirms the target audience, style ("modern and minimalist"), information focus ("concise to focus only on key takeaways"), length ("short, three to five slides"), and core topics.

The merch-launch test produced slides covering an introduction, launch timeline, road map, product lineup, and pricing — tailored to the specified colors with placeholders throughout. Images aren't attached automatically; users click "replace image" and upload files manually. Users can also edit, remove, or add steps in Gemini's presentation plan before generation, and can match new decks to existing ones by attaching them from Drive.

Your AI Workflow's verdict: the five features significantly improve productivity and creativity across these tasks, with the website builder and PDF filler standing out. The pattern across all of them is consistent — generation is fast and visual output is polished, but attaching files, replacing images, and narrowing questions remain human work. As Gemini's embeds and integrations mature, expect those residual manual steps to shrink release by release.