Canaryusers

ai.canaryusers由 canaryusers.ai 發布官方發布者v1.2.2更新於 Oct 24, 2026

A flock of AI users tests your deployed app and reports where real people get stuck, with fixes.

已驗證Streamable HTTP可網頁執行瀏覽器自動化開發者工具

概覽

AI 產生的概覽

讓助理對已部署的網頁應用程式 URL 執行 AI 使用者 UX 掃描，並回傳依優先順序排列、可直接修復的問題。

可以做什麼
CanaryUsers 會讓行為多樣的 AI 使用者角色瀏覽已部署的網頁，並回報使用者卡住的地方，例如令人困惑的 UX、行動版版面損壞、走不通的轉換路徑和視覺異常。canary_scan 工具對 URL 執行掃描，支援 quick（靜態檢查，免費）或 deep（在桌機與行動裝置上視覺化呈現並點擊關鍵流程，消耗額度）模式。list_recent_scans 列出最近的掃描及其 id、URL、分數、等級、狀態和日期，get_report_markdown 以 Markdown 取得完整報告。
適用情境
適合在發布或部署 UI 變更之後使用，因為它掃描的是已部署的結果而非原始碼。也適合上線前對首次使用體驗做壓力測試、檢查註冊、結帳、引導或定價等關鍵流程，以及排查疑似行動版損壞或轉換流失。
執行需求
遠端 Streamable HTTP 端點 CanaryUsers API 權杖，透過 Authorization 標頭以 Bearer 權杖形式傳送；權杖需在 CanaryUsers 儀表板的 CI & API 中開啟 MCP 後產生。需要能存取該端點以及被掃描的 URL。
安裝前請注意
Authorization 標頭攜帶機密 API 權杖，請勿放入共用設定。deep 掃描會消耗額度，額度模型大致為每掃描一個頁面消耗一個額度，因此可能產生費用；quick 掃描免費。掃描針對已部署的 URL，服務會代你呈現該頁面及其流程。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Canaryusers，將其新增到工作區。
  2. 連接 Authorization token 憑證。它們會被私密儲存，絕不會在公開頁面上展示。
  3. 為需要使用其工具的對話啟用該服務。

Web executable，透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/ai.canaryusers%2Fcanaryusers/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "canaryusers": {
      "type": "http",
      "url": "https://www.canaryusers.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer <YOUR_TOKEN>"
      }
    }
  }
}

README

CanaryUsers — MCP Server

[Official MCP Registry] [Smithery] [Website]

Give your AI coding assistant eyes for UX: a flock of AI users tests your deployed app and reports where real people would bail — each with a concrete fix.

CanaryUsers sends a flock of behaviorally‑diverse AI users through your deployed web app (a live or preview URL) and reports exactly where real people get stuck — confusing UX, broken mobile layouts, dead‑end conversion paths, and visual glitches — each with a concrete, AI‑ready fix you can apply in the same chat.

This is the public home for the CanaryUsers remote MCP server: manifest, install instructions, and docs. The product itself is hosted (closed source) — there's nothing to clone or run; you connect to the live endpoint. See docs/WHY.md for why this repo exists and docs/HOW-IT-WORKS.md for how the server works.

  • Registry name: ai.canaryusers/canaryusers
  • Endpoint (remote, Streamable HTTP): https://www.canaryusers.ai/api/mcp
  • Auth: Authorization: Bearer <your CanaryUsers token>
  • Setup / get a token: https://www.canaryusers.ai/mcp

What it does

You point it at a URL; a flock of AI personas actually renders and explores the page like real users, then hands back prioritized, fix‑ready findings — not a vague score. Two modes:

  • Quick (default) — a fast static check. Free.
  • Deep — renders the page, inspects it visually on desktop + mobile (catches mobile breakage and layout issues), and clicks through key flows like signup / checkout / onboarding (~60–90s, uses credits).

When to use it

  • Right after you ship or deploy a UI change (it scans the deployed result, not your source).
  • Before a launch, to pressure‑test the first‑run experience.
  • To check a specific high‑stakes flow: signup, checkout, onboarding, pricing, or a landing page.
  • When you suspect mobile breakage or a conversion drop‑off you can't see in analytics.

Why not Lighthouse or analytics?

Lighthouse grades your HTML; analytics charts what already happened. CanaryUsers sends users and tells you, in plain language, where they'd give up and why — before real ones do.

Tools

ToolWhat it does
canary_scanRun a UX scan on a deployed URL. depth: quick (free, default) or deep (visual + flow, uses a credit). Returns AI‑ready findings.
list_recent_scansList your recent scans (id, URL, score, grade, status, date).
get_report_markdownFetch the full report for a scan id as Markdown.

Install

You need a CanaryUsers token: canaryusers.ai/dashboard → CI & API → toggle MCP on → copy your token.

Generic (Cursor ~/.cursor/mcp.json, Windsurf, VS Code, Claude Desktop):

json
{  "mcpServers": {    "canaryusers": {      "url": "https://www.canaryusers.ai/api/mcp",      "headers": {        "Authorization": "Bearer YOUR_CANARYUSERS_TOKEN"      }    }  }}

Claude Code (CLI):

bash
claude mcp add --transport http canaryusers https://www.canaryusers.ai/api/mcp \  --header "Authorization: Bearer YOUR_CANARYUSERS_TOKEN"

Then just ask your assistant: "Scan my app at https://… with CanaryUsers."

Pricing

Quick scans are free. Deep scans use credits (the credit model is roughly 1 credit = 1 scanned page). Plans and monthly allotments change over time, so this repo intentionally does not hardcode the numbers — see the live pricing for current plans and limits:

➡️ https://www.canaryusers.ai/#pricing (single source of truth)

Links

License

MIT — applies to the contents of this metadata/docs repo. The CanaryUsers service itself is a separate hosted product.

來源：README.md，提交 d07636f

工具

3

canary_scan

Run a CanaryUsers UX scan on a DEPLOYED URL (your live or preview app — not source code). A flock of AI personas evaluates the page and reports where real users would get stuck, with concrete fixes. Returns AI-ready findings you can act on immediately. Use depth='deep' for the thorough scan that renders the page, checks it VISUALLY on desktop + mobile (catches mobile breakage and layout issues), and clicks through key flows like signup/checkout (slower, ~60-90s, uses one credit); depth='quick' (default) is a fast static check that does NOT see mobile or visual issues — use 'deep' when the user mentions mobile, layout, or visual problems. IMPORTANT: if this returns status 'running' with a scanId, the findings are not ready yet — wait ~30s, then call get_report_markdown(scanId), repeating until it returns the report. Always fetch and present the findings before stopping, then offer to fix the top issues.

輸入參數url必填depth

list_recent_scans

List the most recent CanaryUsers UX scans for your account (id, URL, CanaryScore, grade, status, date). Use this to find a scan to dig into.

輸入參數limit

get_report_markdown

Get one scan's full report as AI-ready markdown: the findings ranked by severity, why each matters, and how to fix it. Pass a scanId from list_recent_scans.

輸入參數scanId必填

版本歷史

1
  1. v1.2.2最新Sep 16, 2026

評分

已滿足適用檢查項的 100%

該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分