3 New AI Tools Worth Trying This Week

Stitch by Google : From Prompt to Figma in Seconds

Google just launched Stitch, a new AI-powered design assistant (in beta). It lets you generate responsive UI designs for mobile or web apps from plain prompts, no code or design skills required.

To test Stitch, we asked it to generate a mobile UI for a handmade ceramics marketplace. You can describe your project directly in the prompt box and choose between Mobile or Web layouts.

Before generating, Stitch lets you pick between two modes:

• Standard Mode (powered by Gemini 2.5 Flash): Fast and lightweight, ideal for quick iterations. You can export results directly to Figma or code, and customize themes right after generation.

• Experimental Mode (powered by Gemini 2.5 Pro): Focuses on high-fidelity designs based on your prompt — and even accepts visual inspiration through uploaded images. However, it currently doesn’t support export to Figma.

After a few seconds, here’s what Stitch generated.

On the left, we see a detailed breakdown of the design structure, followed by the actual UI frames generated from the prompt.

Each frame is interactive: you can edit, copy to Figma, or simply like/dislike to refine future outputs.

There’s also a theme editor on the right, allowing you to tweak colors, corner radius, and typography in real time.

Notably, the result includes fully structured UI components, with clean layout, placeholder content, and a polished aesthetic, ideal for fast iteration or design handoff.

When clicking on any frame, Stitch opens a detailed view of the screen. From there, you can copy the code.

11.ai by ElevenLabs : Your Personalized Voice AI Agent

Eleven is a conversational voice assistant you can customize like your own Jarvis. Built by ElevenLabs, it lets you start voice sessions where the agent can interact with your apps in real-time.

To go a bit deeper into what Eleven can do, the assistant already supports native integrations with popular tools like Google Calendar, Notion, Perplexity, Slack, and more.

That means you can schedule or delete calendar events, create and edit Notion pages, or run live voice-powered web searches, all in real time, just by speaking.

You can fully configure your Eleven agent to match your preferences. Start by naming it, then choose a voice model and language, powered by ElevenLabs’ high-quality voices.

Beyond that, you can define custom behaviors using prompt-based instructions, and even set up Custom MCP Servers to connect apps that don’t yet have native integrations.

It takes a bit of technical setup (tools like Cursor can help), but this flexibility opens the door to building your own voice-powered workflows. Lots of potential here for what’s still an early-stage project.

Gemini CLI by Google : AI Agents for the Terminal

This open-source tool allows you to interact with AI locally: ask questions, generate or edit files, run commands, debug code, and more, all within your development environment.

It integrates seamlessly with your existing tools and workflows, running fully on your machine without sending code to the cloud.

Use cases:

• Auto-generate and refactor code in real-time

• Debug local projects using conversational prompts

• Automate terminal workflows with natural language

Official Blog Post

Feel free to share Be 1% Better Every Monday

Be 1% Better, Every Monday.