基准评分
Dyad 在 Agent 就绪度与 AI 可见性上的得分 AI 就绪度和 GEO Score 是 VibeLaunch 在提交后生成的平台评估。
决策摘要
Developers building web applications with AI assistance
AI-assisted web app development with local execution
适合
- Developers who want AI-assisted coding without vendor lock-in
- Rapid prototyping of JavaScript web apps
- Teams that prefer BYO AI API keys
注意
- Testing panel is experimental and not production-grade
- Requires separate AI API key setup before use
- Limited to local development — no browser-based IDE
概述
概览\nDyad 是一款革命性的免费、本地且开源的 AI 应用构建器,旨在赋能创作者,无需任何编程专业知识即可进行开发。它为云端平台提供了一个强大的替代方案,允许您完全在自己的电脑上构建和运行应用程序。这种方式确保了数据隐私、速度,并消除了供应商锁定。Dyad 的构建理念是让开发者真正拥有自己的工具和作品。\n\n## 什么是 Dyad?\nDyad 为在本地开发 AI 驱动的应用程序提供了一个无缝的环境。它集成了任何 AI 模型,包括 Gemini、GPT-5 和 Claude 等热门选项,并支持通过 Ollama 在本地运行模型以实现最大程度的隐私保护。其 Supabase 集成功能允许创建全栈应用程序,使得在 Dyad 内部构建完整的最小可行产品 (MVP) 成为可能。\n\n## 核心优势\n- 本地与隐私:您的所有代码和数据都保留在您的机器上,确保完整的隐私和安全。\n- 开源:受益于透明、社区驱动的开发过程,并为其成长做出贡献。\n- AI 中立:使用您喜欢的任何 AI 模型,而不受特定供应商的束缚。\n- 无锁定:您的工具、您的代码、您的工作流。Dyad 可以与您现有的开发环境集成。\n- 免费使用:Dyad 核心应用程序完全免费,让每个人都能触及强大的 AI 应用构建能力。\n\n## 主要特性\nDyad 拥有多项旨在提供无摩擦开发体验的功能:\n- 无限应用构建:无限制地创建无限数量的 AI 应用。\n- Supabase 集成:轻松整合身份验证、数据库和服务器函数,进行全栈开发。\n- 本地模型支持:使用 Ollama 直接在您的机器上运行 AI 模型,增强隐私性。\n- 快速流畅的 UX:在构建过程中享受即时反馈和实时预览。\n- 无摩擦本地开发:避免在连接云端和本地环境时经常出现的常见问题。\n\n## 谁应该使用它\nDyad 是开发者、设计师、爱好者和企业家的理想选择,他们希望快速原型化并构建 AI 驱动的应用程序,而无需面对传统编程的复杂性或云端构建器的限制。它非常适合那些优先考虑隐私、��������益和对开发过程拥有控制权的用户。
评价 (0)
还没有评价。成为第一个评价的人!
评分构成
编辑评分由哪些维度构成,每项附判断依据。 AI 就绪度和 GEO Score 是 VibeLaunch 在提交后生成的平台评估。
Information quality
Documentation is clear and practical but entirely vendor-sourced. No third-party validation or independent benchmarks available.
Official docs provide step-by-step quickstart, release notes, and deployment guides. All claims originate from dyad.sh.
Ease of use
Quickstart guide is straightforward, but setup requires downloading the app and configuring separate AI API keys before any value is delivered.
Quickstart walks through download → install → API key setup → first app. GitHub-to-Vercel deployment is described as requiring minimal configuration.
Feature depth
Core AI app building is solid. Testing panel is directionally valuable but experimental. Multi-model support provides flexibility.
Supports Gemini, Claude, and OpenAI. v1.7.0 adds experimental testing panel and memory optimizations for large apps.
Workflow fit
GitHub + Vercel pipeline aligns with modern frontend deployment standards. Local execution may not suit teams preferring cloud IDEs.
Documented GitHub-to-Vercel workflow with automatic Vite detection. Push-to-deploy model matches industry practice.
Reliability
v1.7.0 specifically targets stability and memory issues, suggesting earlier versions had reliability gaps. Testing panel is explicitly experimental.
Changelog details memory bounding across multiple subsystems. Testing panel documentation recommends continued manual review of important flows.
Value
Completely free and open-source. Users only pay for their own AI API usage. No platform fees or usage limits.
Documentation states Dyad is free and open-source. AI costs are user-managed via BYO API keys.
评分反映可查证的产品资料,不代表实际使用效果保证。
Agent 就绪度
评估 Agent 能否通过产品的官方信息理解产品,并重建一条有文档依据的工作流程。
Automated agent-readiness assessment of https://dyad.sh/: 9 of 22 checks verified across 6 fetched pages. No substantial machine interface is documented — agents can understand and cite the product but not operate it. Absent: llms_txt, agent_tooling_artifacts, api_reference, request_examples, response_examples, error_documentation.
就绪度维度
| 评估维度 | 得分 |
|---|---|
| 文档质量 | 65 |
| 执行结果可验证性 | 0 |
| 机器接口 | 10 |
| 项目定位清晰度 | 100 |
| 资源可发现性 | 75 |
| 工作流完整度 | 80 |
对 Agent 有帮助的部分
- docs: verified during this run
- sitemap: verified during this run
- quickstart: verified during this run
- authentication: verified during this run
- rate limits: verified during this run
- version information: verified during this run
Agent 受阻的部分
- llms.txt is absent (HTTP probe during this run).
- No agent instruction files, code-distribution commands, or named slash-command skills found across fetched pages.
- No api reference signal matched across 6 fetched pages.
- No request examples signal matched across 6 fetched pages.
- No response examples signal matched across 6 fetched pages.
- No error documentation signal matched across 6 fetched pages.
| 检查项 | 状态 | 详情 |
|---|---|---|
| 理解产品2/5 已核验 | ||
| 产品文档 | 已核验 | Developer/documentation pages reachable from the entry page (e.g. https://www.dyad.sh/docs). |
| 快速开始 | 已核验 | Probe matched on https://www.dyad.sh/docs: /quick ?start|getting started|in (five|5)/. |
| API 参考 | 未在本次官方来源链中找到 | |
| 请求示例 | 未在本次官方来源链中找到 | |
| 响应示例 | 未在本次官方来源链中找到 | |
| 连接接口2/4 已核验 | ||
| SDK | 未在本次官方来源链中找到 | |
| MCP 接口 | 已核验 | Probe matched on https://www.dyad.sh/docs/releases/1.13.0-beta.1: /model context protocol|\bmcp\b(?!-)/. |
| Webhooks | 未在本次官方来源链中找到 | |
| 认证文档 | 已核验 | Probe matched on https://www.dyad.sh/docs/getting-started/quickstart: /api key|bearer|oauth|access token|authen/. |
| 执行工作流0/6 已核验 | ||
| 命令行工具 | 未在本次官方来源链中找到 | |
| 非交互式命令 | 不适用于该产品 | No CLI was found to evaluate for this property. |
| 命令行结构化输出 | 不适用于该产品 | No CLI was found to evaluate for this property. |
| 结构化导入与导出 | 未在本次官方来源链中找到 | |
| 成功状态验证 | 未在本次官方来源链中找到 | |
| 智能体工具产物 | 未在本次官方来源链中找到 | |
| 维护与排错3/4 已核验 | ||
| 错误文档 | 未在本次官方来源链中找到 | |
| 速率限制 | 已核验 | Probe matched on https://www.dyad.sh/docs: /rate limit|429|throttl|requests per (sec/. |
| 版本信息 | 已核验 | Probe matched on https://www.dyad.sh/docs: /api version|versioning|backward compat/. |
| 更新日志 | 已核验 | Probe matched on https://www.dyad.sh/docs/releases/1.13.0-beta.1: /changelog|release notes|what'?s new/. |
| 发现与验证2/3 已核验 | ||
| llms.txt | 未在本次官方来源链中找到 | |
| 站点地图 | 已核验 | sitemap.xml reachable and lists site pages. |
| 智能体原生定位 | 已核验 | Agent-native positioning with a concrete operational path: "The release notes describe a concrete agentic loop in Build mode, which is an operational path for AI agents to work on apps.". |
官方证据
审计信息
- 评测时间
- 2026年8月30日
- 评测基准
- agent-readiness-v1
- 读取页面
- 6
- 来源深度
- 1
本审计从一个入口 URL 及其经过验证的官方来源链评估文档所支持的可操作性。AIGCLIST 未注册、登录、购买、执行或测试该产品的运行可靠性。
证据核查
关于该工具的公开声明,每条均标注核验状态与引用来源。
docs已验证4www.dyad.sh已验证核验于 2026年8月30日
Dyad is a free, open-source AI app builder that runs locally on the user's computer.
Dyad is positioned as an alternative to cloud-based AI app builders including v0, Lovable, and Bolt.new, emphasizing local execution and no lock-in.
Dyad is open-source, and users can collaborate on the project or host private projects via GitHub.
A documentation surface is reachable at https://dyad.sh/docs.
https://www.dyad.sh/docsdocs/getting-started/quickstart已验证4www.dyad.sh已验证核验于 2026年8月30日
Dyad supports AI models via user-provided API keys, including Google Gemini, Anthropic Claude, and OpenAI.
The vendor recommends starting with Google Gemini's free tier, citing approximately 250 messages per day on Gemini 2.5 Flash.
Users must download and install the Dyad desktop application to get started.
A quick-start / agent-skills documentation page is reachable at https://www.dyad.sh/docs/getting-started/quickstart.
https://www.dyad.sh/docs/getting-started/quickstartdocs/getting-started/publishing-your-app已验证4www.dyad.sh已验证核验于 2026年8月30日
Dyad apps can be deployed by pushing code to GitHub and importing the repository into Vercel, which auto-detects the framework (e.g., Vite) and handles deployment with minimal configuration.
Vercel is recommended as an excellent hosting option for Dyad-built JavaScript apps due to its easy setup and free plan.
Dyad is open-source, and users can collaborate on the project or host private projects via GitHub.
A quick-start / agent-skills documentation page is reachable at https://www.dyad.sh/docs/getting-started/publishing-your-app.
https://www.dyad.sh/docs/getting-started/publishing-your-appdocs/releases/1.7.0已验证4www.dyad.sh已验证核验于 2026年7月15日
Dyad v1.7.0 includes an experimental Testing panel for AI-generated end-to-end testing of user flows.
The Testing panel uses AI to help create end-to-end tests for important user flows, making it easier to verify app behavior after changes.
The Testing panel is explicitly experimental and should be used as an aid alongside continued manual review of important flows before publishing.
Dyad v1.7.0 improved stability and performance for large apps and bounded memory usage across chat attachments, voice-to-text recordings, preview console logs, and child-process output.
https://www.dyad.sh/docs/releases/1.7.01.13.0 Beta 1 | Dyad已验证2dyad.sh已验证核验于 2026年8月30日
A documentation surface is reachable at https://dyad.sh/docs/releases/1.13.0-beta.1.
Agent-native positioning with a concrete operational path: "The release notes describe a concrete agentic loop in Build mode, which is an operational path for AI agents to work on apps.".
https://www.dyad.sh/docs/releases/1.13.0-beta.1Dyad – Free, Local, Open-Source AI App Builder已验证1dyad.sh已验证核验于 2026年8月30日
The entry page was fetched and analyzed for machine-interface signals (title, headings, developer links, keyword probes).
https://www.dyad.sh/https://www.dyad.sh/sitemap.xml已验证1dyad.sh已验证核验于 2026年8月30日
sitemap.xml is reachable and lists site pages.
https://www.dyad.sh/sitemap.xmlhttps://release-assets.githubusercontent.com/github-production-release-asset/964395174/aeb85193-e78b-4893-839f-6985cd0ceb51?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-08-30T13%3A22%3A54Z&rscd=attachmen已验证1dyad.sh已验证核验于 2026年8月30日
An API documentation surface is reachable at https://api.dyad.sh/v1/trusted-assets/v1.12.0/dyad-darwin-arm64-1.12.0.zip.
https://release-assets.githubusercontent.com/github-production-release-asset/964395174/aeb85193-e78b-4893-839f-6985cd0ceb51?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-08-30T13%3A22%3A54Z&rscd=attachment%3B+filename%3Ddyad-darwin-arm64-1.12.0.zip&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2026-08-30T12%3A22%3A23Z&ske=2026-08-30T13%3A22%3A54Z&sks=b&skv=2018-11-09&sig=RRoUwmAfcYYT1PMcdm3t7ROMy14n2PYSngz6Zn6bxx8%3D&jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc4ODA5NjUwNSwibmJmIjoxNzg4MDkyOTA1LCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.WUcnm8uXJJ6shSIN3ypqvoLoVDCdotItT_pfRWBSxZU&response-content-disposition=attachment%3B%20filename%3Ddyad-darwin-arm64-1.12.0.zip&response-content-type=application%2Foctet-stream决策核对台
在依赖该产品或访问官网前,最值得先确认的问题。
Dyad is a free, open-source AI app builder that runs locally on your computer. It helps developers build web applications using AI models like Gemini, Claude, and OpenAI through bring-your-own API keys.
Unlike cloud-based alternatives, Dyad runs entirely on your computer, giving you full control over your code and eliminating vendor lock-in. You can deploy your apps to standard cloud providers like Vercel via GitHub.
Dyad supports Google Gemini, Anthropic Claude, and OpenAI models. Users provide their own API keys, and Gemini's free tier (approximately 250 messages per day on Gemini 2.5 Flash) is recommended for getting started.
Push your code to GitHub, then import the repository into Vercel. Vercel automatically detects your framework (e.g., Vite) and handles deployment with minimal configuration. Environmental variables for API integrations can be added during or after deployment.
Yes, Dyad v1.7.0 includes an experimental Testing panel that uses AI to create end-to-end tests for important user flows. However, the vendor recommends using it as an aid alongside manual review — it is not yet a replacement for thorough QA.
请在官网核验
继续探索
相近任务的不同路径
这些工具以带有明确编辑理由的替代关系关联到当前产品。
MonkeyCode AI
Another AI app builder — cloud-based alternative for developers who prefer a browser IDE over local installation
查看档案Low Code Platforms Directory (低代码平台目录)
Broader category of low-code platforms for comparison with Dyad's AI-assisted coding and local-first approach
查看档案更多 AI 应用构建器 工具
与当前产品共享主要分类的已发布工具。它们用于继续发现,并非编辑对比结论。
