Launching today

SpokeUI

Launching today

Point. Speak. Ship.

3 followers

Point. Speak. Ship.

3 followers

Visit website

SpokeUI is a local desktop workspace for building and refining web interfaces with your voice. Select an element in the live preview, describe the outcome, and let Codex or Claude Code edit the real source. Review the rendered result, changed files, and before/after state before you accept it. Open an existing project or start a React + Vite app from a template. Available for macOS, Windows, and Linux.

Free

Launch tags:Design Tools•Open Source•Developer Tools

Launch Team

Framer AI AgentsDesign and publish professional sites with AI

Promoted

OpenAI

GPT-6 Astra Challenge

What real task does your product handle with GPT-6 Astra?

Maker

GPT-6 Astra helped build SpokeUI by reasoning through its Electron architecture, live browser integration, voice workflow, source editing, debugging tools, and change-review experience. Inside SpokeUI, Astra handles the real task of converting a user’s spoken, visually grounded request into working frontend code. A user can select an element in their running website, say what should change, and Astra receives the relevant DOM, styles, geometry, errors, and project context to locate the correct source files, make the edit, and verify the result in the live preview. This helps developers build, refine, and debug interfaces without manually translating visual problems into selectors, file paths, and long technical prompts.

Report

23h ago

Maker

📌

Hey Product Hunt! I’m Vasu, the maker of SpokeUI. I built SpokeUI because interface feedback usually loses the thing that matters most: the interface itself. You see a button that feels cramped, a layout that breaks at tablet width, or an error state that makes no sense... then you have to translate that visual problem into selectors, file paths, screenshots, and a long prompt. SpokeUI keeps that context attached. You select the element in the running app, say what should change, and a local coding agent edits the real source. The preview updates live, and you can inspect the changed files and rendered before/after state before accepting the result. The current release supports: 1. Existing web projects and new React + Vite starters 2. Codex and Claude Code as local agent runtimes 3. Voice requests through AssemblyAI transcription 4. Element selection, responsive preview, inspector context, and visual debugging 5. Source-backed change review with accept and reject It’s open source, free, and available now for macOS, Windows, and Linux. Release builds are currently unsigned, so your OS may show the usual first-run warning. I’d especially value feedback from developers and engineers: where does the select → speak → review loop save you time, and where does it still get in your way?

Report

2d ago

OpenAI

GPT-6 Astra Challenge

What real task does your product handle with GPT-6 Astra?

Maker

Report

23h ago

Maker

📌

Report

2d ago