InertiaRSS Track and read blogs, news, and tech you care about
Read Original Open in InertiaRSS

Recommended Feeds

博客园 - 司徒正美
V
V2EX
T
Tailwind CSS Blog
有赞技术团队
有赞技术团队
aimingoo的专栏
aimingoo的专栏
Apple Machine Learning Research
Apple Machine Learning Research
IT之家
IT之家
Blog — PlanetScale
Blog — PlanetScale
A
About on SuperTechFans
月光博客
月光博客
T
The Blog of Author Tim Ferriss
宝玉的分享
宝玉的分享
Martin Fowler
Martin Fowler
博客园 - 聂微东
The GitHub Blog
The GitHub Blog
V
Visual Studio Blog
WordPress大学
WordPress大学
酷 壳 – CoolShell
酷 壳 – CoolShell
Engineering at Meta
Engineering at Meta
GbyAI
GbyAI

阮一峰的网络日志

科技爱好者周刊(第 396 期):互联网通信的替代方案 科技爱好者周刊(第 396 期):互联网通信的替代方案 - 阮一峰的网络日志 科技爱好者周刊(第 395 期):软件开发的第三种方式 科技爱好者周刊(第 395 期):软件开发的第三种方式 - 阮一峰的网络日志 科技爱好者周刊(第 393 期):脑腐状态 科技爱好者周刊(第 392 期):axios 投毒与好莱坞式骗术 科技爱好者周刊(第 391 期):AI 的贫富分化 科技爱好者周刊(第 390 期):没有语料,大模型就是智障 套壳中国大模型撑起500亿美元估值?扒一扒 Cursor 的"套壳"疑云 科技爱好者周刊(第 389 期):未来如何招聘程序员 科技爱好者周刊(第 388 期):测试是新的护城河 零安装的"云养虾":ArkClaw 使用指南 科技爱好者周刊(第 387 期):你是领先的 科技爱好者周刊(第 386 期):当外卖员接入 AI 字节全家桶 Seed 2.0 + TRAE 玩转 Skill 科技爱好者周刊(第 385 期):马斯克害怕中国车企吗? 智谱旗舰 GLM-5 实测:对比 Opus 4.6 和 GPT-5.3-Codex 科技爱好者周刊(第 384 期):为什么软件股下跌 科技爱好者周刊(第 383 期):你是第几级 AI 编程 Kimi 的一体化,Manus 的分层 科技爱好者周刊(第 382 期):独立软件的黄昏 AI native Workspace 也许是智能体的下一阶段 科技爱好者周刊(第 381 期):中国 AI 大模型领导者在想什么 科技爱好者周刊(第 380 期):为什么人们拥抱"不对称收益" 科技爱好者周刊(第 379 期):《硅谷钢铁侠》摘录 我如何用 AI 处理历史遗留代码:MiniMax M2.1 升级体验 科技爱好者周刊(第 378 期):预测是新的互联网热点 科技爱好者周刊(第 377 期):14万美元的贫困线 科技爱好者周刊(第 376 期):太空数据中心的争议 科技爱好者周刊(第 375 期):一扇门的 Bug 终于有人做了 Subagent,TRAE 国内版 SOLO 模式来了 科技爱好者周刊(第 374 期):6GHz 的问题 VS Code 使用国产大模型 MiniMax M2 教程 科技爱好者周刊(第 373 期):数据模型是新产品的核心 国产大模型接入 Claude Code 教程:以 Doubao-Seed-Code 为例 科技爱好者周刊(第 372 期):软件界面如何设计 大模型比拼:MiniMax M2 vs GLM 4.6 vs Claude Sonnet 4.5 科技爱好者周刊(第 371 期):一个乐观主义者的专访 科技爱好者周刊(第 370 期):正确的代码高亮 错误处理:异常好于状态码 科技爱好者周刊(第 369 期):Tim 与罗永浩的对谈 科技爱好者周刊(第 368 期):不要这样管理软件团队 一天之内,智谱和 Anthropic 都发了最强编程模型 科技爱好者周刊(第 367 期):Nano Banana 的几个妙用 科技爱好者周刊(第 366 期):旧金山疯狂的 AI 广告 科技爱好者周刊(第 365 期):流量变现正在崩塌 科技爱好者周刊(第 364 期):最难还原的魔方 科技爱好者周刊(第 363 期):最好懂的神经网络解释 科技爱好者周刊(第 362 期):GitHub 工程师谈系统设计 科技爱好者周刊(第 361 期):暗网 Tor 安全吗?
Button Space web design, is it challenging V0?
阮一峰 · 2025-07-24 · via 阮一峰的网络日志

I.

The Button (coze.cn) is probably the most difficult product to define under ByteDance. Its positioning is an AI agent platform, capable of developing/running various AI applications

. This means that anything AI can do, it can also do. However, AI is constantly evolving, so the form of The Button is also constantly changing.

Initially, The Button was a development platform for programmers. It has

a workflow mode, where you drag and drop nodes on a graphical interface to define AI workflows and generate applications. However, soon after, it also launched the "

The Button Space". is designed for regular users, using natural language to call its built-in various agents.

I have written reviews and used it as an office assistant to complete various tasks (PPT generation, travel map planning, writing research reports, etc.), it's completely user-friendly.

But I didn't expect that yesterday I found that Kousen Space has added web design, and it has become a UI tool, similar to having an integrated V0.

II.

V0 is the earliest and currently the most famous AI tool for web design/creation. Users describe what website they want, and it builds it.

This time, Kousen Space has introduced this feature, adding another product form for itself—web design tool.

Now, you can completely develop web pages inside the buttons. Moreover, it's a free service, with an Android client (released) and an iOS client (upcoming), allowing operation on mobile phones.

This sounds very appealing. I tried it out, such as screenshot restoration and website generation, especially restoring Figma design drafts, to see if it actually works.

Below are my trial experiences.

III. Screenshot Restoration

I first tested the screenshot restoration, uploading a webpage screenshot and letting it generate the page.

The webpage design feature of the buttons must have the "Webpage" button in the buttons space enabled.

I uploaded the homepage screenshot of airbnb (below), and entered "Restore this screenshot as a webpage".

The system prompts that it needs 3 to 5 minutes of thinking, and the code is generated by scrolling to the right side of the page.

After a few minutes, the code generation is complete. Switch to the preview page, and below is the page it generated.

This is the first generated result,未经调整, I think the还原度 is quite good, the page structure and each area are correctly replicated.

I have shared it, this page and the entire AI dialogue, everyone clicks in, can see the clear effect.

It's worth mentioning that the button generates a React + Tailwind project, which can be entirely downloaded to the local machine and run with npm. This way,后期的手工处理 is quite convenient.

Four、Figma design稿还原

The real killer feature of the button space is generating web pages from Figma design drafts.

The specific method is to first click the Figma button below in the prompt box.

The webpage will prompt you to enter the Figma link.

How do I get this link? In the workspace on the Figma website, select a page, open the right-click menu "Copy/Paste/Copy Selection Link" (see image below), and that's it.

I chose a Figma official website.Example fileSend the link to Koushi Space. Let everyone see Koushi.Fidelity.

Basically, it's very close to the design稿, with a high degree of accuracy. So,If there are original design files, AI tools can completely replace manual development.

This belongs to advanced features, and I haven't seen other AI tools in China that offer this; it also requires a paid subscription in V0.

5. Visual Editing

After preview generation, Koushu Space provides visual editing, making modifications quite convenient.

First, on the preview interface, click the "Edit" button.

Then, select a page element, and an AI dialog box will pop up where you can edit using natural language.

After selecting the button, I asked the AI to change the button's background color to green (as shown in the upper image). Upon submission, the code will be regenerated (as shown in the lower image).

6. Zero-Foundation Website Generation

Below, I directly describe the requirements in natural language to the AI to generate a website.

I asked it to generate a to-do (Todos) application.

"Create a to-do list app with a blue theme and implement editing functionality, allowing you to edit list tasks."

The Button Space then automatically completes the page design, code writing, server deployment, and finally providesThe effect below.

This page can also be deployed asAccessible linksOpen to external users.

I think it's quite good, and there were no errors in various operations (adding tasks, editing tasks, deleting tasks). It runs with a single generation, and all the necessary buttons are present.

This really exceeded my expectations. Is it because the buttons have been specially trained for common tasks?

Theoretically, website generation doesn't require professional knowledge, and users with zero experience can create entire websites. In practice, this is also true. Even if errors occur, they can be handed over to AI for processing. However, this is only limited to simple interactive websites. If JavaScript scripts are needed to handle complex interactions, it's likely that errors may arise that AI cannot resolve.

7. Summary

That's the initial impression of using it.

I feel thatThe web design feature of Button Space, although just launched, has a very good completion level..

Its three key features—design draft/screenshot restoration, visual editing, website generation—are quite useful, and the generated quality is perfectly acceptable. Plus, it's free, which I think makes it very competitive.

It also has some usage restrictions.

(1) No command line window。If you don't need the UI, just let it generate backend code, it doesn't provide a command line, making it impossible to test.

(2) Not integrated with GitHub 。It cannot sync code directly with GitHub; you have to download it locally first and then push it manually to GitHub.

(3) Not suitable for complex state management 。Its strength lies in UI generation, making it unsuitable for complex state management and interaction logic.

Overall, this update has made Button Space more useful. In addition to the earlier chat and office functions, it now includes web design, making it more practical. Common UI development tasks (generating web pages and components, completing prototype design) can definitely be considered for it.

(End)