Carryo

Carryo is the easiest way to publish what you create with AI as a real, live website. Create a presentation, proposal, report, landing page, dashboard, invitation, portfolio, or simple website in chat, then say 'publish this as a website' - Carryo gives you a shareable live link in seconds. No hosting setup, domain, or deployment pipeline needed. Keep editing the published website from your AI chat while preserving the same URL, add images, browse earlier versions, list your websites, or take one offline. Want to review before going live? Save a private website draft and publish it from Carryo when you're ready. For shared work, explicitly choose a Carryo team so authorized members can publish and maintain the same websites from their own connected accounts. After publishing or editing, open Carryo to manage the URL, password, view count, edit history, and team settings.

该来源不提供完整文件导出（国内平台多为平台内托管），仅存元数据与原链

模型生成摘要（rules/v1 · 2026-09-27 16:10）：Carryo is the easiest way to publish what you create with AI as a real, live website
这是模型对公开材料的总结，不是官方声明，请以原链内容为准。

接入信息

传输形态: http
鉴权方式: 鉴权未知
端点: https://publish-a-website.run.tools

鉴权方式未标注，请核对官方文档后再接入——不要直接使用以下片段

{
  "mcpServers": {
    "Carryo": {
      "url": "https://publish-a-website.run.tools"
    }
  }
}

能力清单

工具	说明
account.check_publishing	Use this when the user asks whether they can publish, how many active or new Personal websites remain, when the publishing allowance resets, how inactivity works, or whether a custom URL ending, password-page branding, or badge setting can be used. Checks whether the user can publish another Carryo website and shows remaining capacity. Returns only publishing capabilities, relevant usage, and website policies; it does not return billing or purchase information. Read-only; changes nothing.
website.publish	Use this when the user wants to publish, share, host, or put online something they created in this conversation, including a presentation, proposal, report, landing page, dashboard, invitation, portfolio, or simple website. Publishes the AI-chat creation as a real, live website and returns a shareable link plus a Carryo management link for the URL, password, views, and edit history. Publish to Personal by default. Only pass workspaceId when the user explicitly chose a team resolved with team.list; never infer one. A new Personal website uses the account creation allowance; do not create a replacement website to bypass an edit or creation limit. If a limit is reached, report it factually and direct the user to the returned Carryo management location without giving purchase instructions. Ask whether the user wants a custom URL ending and pass it as customSlug when provided; otherwise omit customSlug. Images: reference relative paths such as images/photo.jpg in htmlContent and provide matching assets entries with path and a fetchable http(s) sourceUrl; never use base64 or data URLs. Do not use for non-HTML files, general file storage, backend app hosting, websites that collect viewer data, or websites requesting restricted sensitive personal data.
website.save_draft	Use this when the user wants the website to stay private for now, wants restricted access, or wants to review settings before anything goes live. Saves the AI-chat creation as a private Carryo website draft and returns a Carryo management link, not a public website link; the website stays inaccessible until the user publishes it from Carryo. A new Personal draft uses the same creation allowance as a published Personal website; do not create drafts or replacement websites to bypass another limit. Save to Personal by default. Only pass workspaceId when the user explicitly chose a team resolved with team.list; never infer one. Never collect access secrets in ChatGPT. Images follow the same rules as website.publish: matching assets entries with fetchable http(s) sourceUrl, never base64 or data URLs. Do not use for websites that collect viewer data or request restricted sensitive personal data.
website.read	Use this when the user wants to edit a website they already published or drafted with Carryo, or asks what a live link currently contains. Reads the current version before editing so the AI can preserve the existing website. Call this before website.edit. Returns image details, not image files.
website.list_versions	Use this when the user asks about edit history, previous versions, what changed, or wants to restore an older Carryo website. Lists previous website versions and shows what changed. Returns version details only; for normal edits use website.read for the current version instead.
website.read_version	Use this when the user has picked a specific previous Carryo website version to view, compare, or restore. Reads and returns that version of the website. For normal edits use website.read for the current version instead.
website.edit_images	Use this when the user wants to add, swap, or remove images on an existing Carryo website without rewriting the whole website. Adds, replaces, or removes hosted images. Provide path and a fetchable http(s) sourceUrl per image; never read image bytes or use base64 or data URLs. If no fetchable URL exists, stop and ask the user for one. Afterwards call website.edit only if the website must be changed to show the returned image URL.
website.edit	Use this when the user wants to change, fix, improve, or republish a website that already has a Carryo link. Edits the existing website while keeping the same live link. Read the current website with website.read first, then save the revised version here. Team access is resolved securely from linkId; never ask for or supply a team identity. Returns the unchanged live link plus a Carryo management link for the URL, password, views, and edit history. If the edit limit is reached, do not suggest publishing a replacement website as a workaround; report the limit factually and direct the user to Carryo management without giving purchase instructions. Add or replace images via assets entries with path and a fetchable http(s) sourceUrl, never base64 or data URLs, and remove images with removeAssetPaths. Do not use for websites that collect viewer data or request restricted sensitive personal data.
team.list	Use this when the user explicitly asks to publish, save, or list websites for a named Carryo team. Finds the team and returns its workspaceId for the next tool call. Do not guess a team or select one merely because the user belongs to it. Returns only names, roles, and available actions; team creation, invitations, members, and billing are managed in the Carryo web app.
website.list	Use this when the user asks to see, find, or manage what they've published, such as 'show my websites', 'what have I published with Carryo?', or 'find my pricing page', or whenever you need a linkId and the user referred to a website by name instead of ID. Finds the user's live Carryo websites and private drafts. Lists Personal websites by default. Only pass workspaceId when the user explicitly chose a team resolved with team.list; never combine or guess collections. Returns Carryo websites with titles, live links, and status.
website.unpublish	Use this when the user wants to unpublish, take down, disable, or delete a Carryo website or live link. Takes the website offline so its shared link stops working. Viewers lose access immediately, but the user can restore the website in Carryo until the returned deletion date. Confirm with the user before calling.

纠错与举报（发现条目失效、署名有误或涉及侵权？）

提交举报 / 纠错

侵权举报经核验成立后，我们会即时下线该条目并删除已存的内容副本。