Vexa
Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.
作者提供的说明(英文原文)
Deep links to pixel-perfect fake chat & screenshot generators (iMessage, WhatsApp & more).
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"mockscreenshots": {
"command": "npx",
"args": [
"-y",
"@gamercc-apps/mockscreenshots-mcp"
]
}
}
}在终端中运行这条命令
claude mcp add mockscreenshots -- npx -y @gamercc-apps/mockscreenshots-mcp粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"mockscreenshots": {
"command": "npx",
"args": [
"-y",
"@gamercc-apps/mockscreenshots-mcp"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"mockscreenshots": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@gamercc-apps/mockscreenshots-mcp"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月7日
注册名称已通过 GitHub 账户 gamercc-apps 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(0.1.6)于4个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js。
同一分类中有:Vexa, emailmd, MCP WhatsApp Cloud
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.
Render markdown into email-safe HTML, lint drafts for deliverability problems, and preview emails.
MCP server for WhatsApp Cloud API (Meta) — messages, media, and templates via Graph API
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
An MCP server that lets an AI agent compose a fake chat and get a deep link to the matching Mock Screenshots generator — pre-filled and ready to preview and download.
Output is watermarked and clearly fictional, intended for parody, education, design mockups and fiction. It is not for deception — see the ethics policy.
| Tool | What it does |
|---|---|
generate_fake_chat | Compose a conversation (platform, messages, contact, status, device, dark) → returns a rendered, watermarked PNG (inline preview + hosted URL for download/share) plus a deep link to the generator. Supports format: "image" (default, returns preview+URL) or "link" (text-only URLs). |
list_platforms | Lists supported chat apps and their generator URLs. |
list_devices | Lists the iPhone/Android device frames. |
generate_fake_chat input{
"platform": "whatsapp", // imessage | whatsapp | whatsapp-group | instagram | telegram | messenger | snapchat "contact": "Mom", // header name / username / group name "status": "typing…", // header status line (optional) "device": "galaxy-s24", // see list_devices (default iphone-16-pro) "dark": true, // dark mode (optional) "format": "image", // "image" (default) | "link" — see Screenshots section "messages": [ { "text": "you home?", "sender": "them", "time": "19:01", "image": { "data": " ", "mimeType": "image/png", "alt": "A synthetic design mockup" } }, { "text": "5 mins!", "sender": "me", "time": "19:02", "ticks": "read" } ] }
messages[].image is optional and supported only for whatsapp and whatsapp-group, so existing text-only requests remain unchanged. Image-only messages are also supported. Attachments must be self-contained PNG, JPEG, or WebP bytes with a matching file signature; SVG, remote URLs, malformed base64, and files over 2 MB are rejected. Alternative text is limited to 160 characters, trimmed, and defaults to Attached image. The server also rejects more than 100 messages, message text over 4,000 characters, header metadata over 256 characters, and aggregate text/image source data over 6,000 bytes before it decodes attachments or builds duplicate JSON/base64url state. The merged site endpoint limits the complete base64url state to 8,000 characters, so the final encoded-state check remains authoritative and a smaller attachment may still be required after conversation metadata is included.
Use only non-sensitive synthetic or already-public attachments. Attachment bytes are stored as a data URL inside the base64url-encoded JSON state in both the hosted GET /api/render?...&s= URL and the generator edit ?s= URL. Base64url is encoding, not encryption: anyone or any system with either URL can recover the attachment.
Treat every attachment-bearing render or edit URL as sensitive. It may be retained in MCP transcripts, client logs, browser history, proxy/CDN request logs, analytics/referrers, and cache keys. Do not use credentials, private user data, confidential screenshots, or private uploads as fixtures. The MCP's own preview request uses cache: "no-store" and referrerPolicy: "no-referrer" as defense in depth, but those client-side request settings cannot guarantee that the deployed endpoint, intermediaries, or a user-opened URL will not log or cache the full GET URL. No attachment or URL logging is added by this MCP server.
This residual exposure is inherent in the currently deployed URL-state contract. Roll back attachment support by reverting the attachment commit, or avoid messages[].image; existing text-only requests remain compatible and unchanged. A future opaque, expiring server-side state token would require a separate site/API contract and is intentionally not invented here.
Returns a rendered, watermarked screenshot server-side via the site's GET /api/render endpoint (Cloudflare Browser Rendering). Default output (format: "image") includes:
Alternately, use format: "link" for text-only output (just the URLs, no image preview).
Always watermarked: Screenshots include a prominent "FAKE" watermark and cannot be disabled. This ensures they remain clearly fictional and non-deceptive for parody, education, design mockups and fiction — see the ethics policy.
Also returns a URL like https://mockscreenshots.com/fake-whatsapp-chat-generator?s= that opens the generator with the conversation loaded (for preview/tweaking before final export).
npm install
npm start # stdio server
// claude_desktop_config.json (or: claude mcp add)
{ "mcpServers": { "mockscreenshots": { "command": "node", "args": ["/absolute/path/to/mcp/server.mjs"] } } }
Once published to npm you can instead use npx -y mockscreenshots-mcp.
npm publish the mockscreenshots-mcp package.registry.modelcontextprotocol.io using server.json.awesome-mcp-servers.Each listing is a genuine dofollow dev-domain link + agent discovery — the uncontested distribution channel
The MCP server itself stays tiny and stateless: it builds URLs and encodes the conversation into a compact, URL-safe ?s= parameter (also read by the generator on load, src/lib/share.ts), then fetches a preview from the site's /api/render endpoint, which does the actual (always-watermarked) server-side rendering via Cloudflare Browser Rendering. This keeps the server dependency-light and keeps the human in the loop to preview, tweak and export.
The server never fetches a caller-supplied attachment URL. Preview responses are accepted only when they are PNG data from the fixed Mock Screenshots endpoint. The response body is consumed with a 10 MB streaming cap and the reader is cancelled immediately when that cap is crossed, even when Content-Length is absent or inaccurate. Endpoint errors, invalid responses, and timeouts safely fall back to the hosted image/edit links while retaining the watermark, ethics warning, and (when an attachment is present) attachment privacy warning.