Vynix

in.vynix由 vynix.in 發布官方發布者v0.1.4更新於 Oct 27, 2026

Give AI coding agents access to your Vynix visual feedback, bug reports, and AI diagnosis.

已驗證Streamable HTTP可網頁執行效率與工作流開發者工具AI 與機器學習

概覽

AI 產生的概覽

Vynix MCP 讓編碼助理存取 Vynix 的視覺回饋、缺陷回報、截圖、診斷資訊與 GitHub 議題工作流程。

可以做什麼
  • 列出專案、標註、成員與活動紀錄，並讀取單一標註所擷取的頁面、目標元素、DOM 與執行期診斷
  • 檢視標註附帶的區域截圖與留言討論串
  • 讀取已儲存的 AI 診斷，或執行 AI 診斷引擎取得排序後的根因、信心分數、修正建議與可能的原始檔
  • 為指定助理產生可直接貼上的編碼提示詞
  • 從標註建立 GitHub 議題，並依標註或專案列出追蹤議題
  • 更新標註狀態、新增留言，並產生唯讀的公開審閱連結
適用情境
適合團隊在 Vynix 收集視覺缺陷與回饋，並希望編碼助理依據實際擷取的脈絡而非猜測來作業。適用於分流、診斷、提示詞產生以及交接至 GitHub 議題。
執行需求
需要 Vynix 帳號，並在 VYNIX_API_TOKEN 提供個人存取權杖（文件亦提到電子郵件／密碼的備援方式）。可作為遠端 Streamable HTTP 端點使用，或透過 npx 在本機執行，需 Node.js 18+；VYNIX_API_URL 為選用，預設指向 Vynix 網站。需要連線至 Vynix API 的網路，議題相關工具經 Vynix 後端存取 GitHub。
安裝前請注意
並非唯讀：工具會變更標註狀態、張貼會通知團隊的留言、建立 GitHub 議題並產生公開分享連結。diagnose_annotation 會呼叫外部 AI 供應商，消耗 token 且可能產生費用。API 權杖屬於憑證，不應提交至版本庫，建議使用短期權杖。依 README，僅診斷功能會將資料傳送給第三方 AI 供應商。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Vynix，將其新增到工作區。
  2. 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/in.vynix%2Fvynix-mcp/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": {
    "vynix": {
      "type": "http",
      "url": "https://mcp.vynix.in/mcp"
    }
  }
}

README

Vynix MCP Server

[npm version] [npm downloads] [license] [ci] [vynix-mcp MCP server]

Model Context Protocol server for Vynix. It gives coding agents direct access to visual feedback, bug reports, screenshots, diagnostics, comments, and issue workflows so agents can reason from real context instead of guessing.

Why Vynix

  • Feedback with evidence: page metadata, target element, screenshot, console/network context.
  • End-to-end execution: inspect feedback, diagnose, generate coding prompts, create GitHub issues, update status, comment.
  • Agent-safe hints: read-only/idempotent/open-world annotations for better approval behavior in MCP clients.
  • Registry-ready metadata for modern MCP directories.

Architecture

mermaid
flowchart LR  A[MCP Client\nClaude/Cursor/Copilot/VS Code] -->|stdio or streamable-http| B[Vynix MCP Server]  B --> C[Vynix API]  B --> D[GitHub API via Vynix backend]  C --> E[Projects]  C --> F[Annotations]  C --> G[Screenshots and diagnostics]

Features

  • 17 production tools for read and write workflows.
  • Resource catalog for server metadata, tool/prompt/skill references, and contextual summaries.
  • Workflow prompts for QA, release readiness, PM briefings, and engineering planning.
  • Dual transport support: stdio and Streamable HTTP.
  • Auth via API token or email/password refresh flow.

Installation

Node.js 18+ is required.

NPX (recommended)

json
{  "mcpServers": {    "vynix": {      "command": "npx",      "args": ["-y", "@usevynix/mcp-server"],      "env": {        "VYNIX_API_URL": "https://www.vynix.in",        "VYNIX_API_TOKEN": "PASTE_YOUR_TOKEN_HERE"      }    }  }}

npm global install

bash
npm install -g @usevynix/mcp-servervynix-mcp

Docker

bash
docker run --rm -i \  -e VYNIX_API_URL=https://www.vynix.in \  -e VYNIX_API_TOKEN=PASTE_YOUR_TOKEN_HERE \  ghcr.io/usevynix/vynix-mcp:latest

Local development

bash
git clone https://github.com/UseVynix/vynix-mcp.gitcd vynix-mcpnpm installnpm run buildnpm run checknode dist/index.js

Client Configuration

Claude Desktop

Use claude_desktop_config.json:

json
{  "mcpServers": {    "vynix": {      "command": "npx",      "args": ["-y", "@usevynix/mcp-server"],      "env": {        "VYNIX_API_TOKEN": "PASTE_YOUR_TOKEN_HERE"      }    }  }}

Diagnostics are written to stderr; stdout is reserved for the protocol stream.

Cursor

Use ~/.cursor/mcp.json.

json
{  "mcpServers": {    "vynix": {      "command": "npx",      "args": ["-y", "@usevynix/mcp-server"],      "env": {        "VYNIX_API_TOKEN": "PASTE_YOUR_TOKEN_HERE"      }    }  }}

VS Code

Use .vscode/mcp.json with top-level servers:

json
{  "servers": {    "vynix": {      "command": "npx",      "args": ["-y", "@usevynix/mcp-server"],      "env": {        "VYNIX_API_TOKEN": "PASTE_YOUR_TOKEN_HERE"      }    }  }}

Windsurf

Use your Windsurf MCP config file with this server block:

json
{  "mcpServers": {    "vynix": {      "command": "npx",      "args": ["-y", "@usevynix/mcp-server"],      "env": {        "VYNIX_API_TOKEN": "PASTE_YOUR_TOKEN_HERE"      }    }  }}

ChatGPT connectors

For hosted mode, use the Streamable HTTP endpoint:

  • Base URL: https://mcp.vynix.in/mcp
  • OAuth discovery: /.well-known/oauth-authorization-server

Generic mcp.json

See examples/configs/mcp.json.

Authentication

Environment variables:

  • VYNIX_API_URL (optional, default: https://www.vynix.in)
  • VYNIX_API_TOKEN (recommended)
  • VYNIX_API_EMAIL and VYNIX_API_PASSWORD (fallback login mode)
  • VYNIX_MCP_MODE (stdio or http)
  • VYNIX_MCP_HOST, VYNIX_MCP_PORT, VYNIX_MCP_PATH (HTTP mode)

Generate a token from: https://www.vynix.in/mcp

Tools, Prompts, and Resources

Examples

Troubleshooting

  • Not configured error: set VYNIX_API_TOKEN or both VYNIX_API_EMAIL and VYNIX_API_PASSWORD.
  • 401 errors: regenerate token and verify API URL.
  • No tools listed: confirm the MCP config key (mcpServers vs servers) for your client.
  • Hosted mode not reachable: verify VYNIX_MCP_MODE=http and check /health.

FAQ

Does this send data to third-party AI providers?

Only diagnose_annotation can invoke external AI providers through your Vynix workspace configuration.

Is this read-only?

No. It includes read tools and write tools. MCP annotations identify mutating/open-world calls so clients can request confirmation.

Can I self-host?

Yes. Run in stdio mode locally or HTTP mode behind your own infrastructure.

Contributing

See CONTRIBUTING.md. For local validation run:

bash
npm run check

Security

  • Never commit API tokens.
  • Prefer short-lived tokens where possible.
  • See SECURITY.md (create one if your org requires a disclosure policy).

License

MIT

Changelog

CHANGELOG.md

來源：README.md，提交 d60e3db

工具

17

list_projects

List projects

List the Vynix projects you own.

無需輸入參數。

list_annotations

List annotations

List a project’s annotations, optionally filtered by status, type, or priority. Start here with status "open" to see what needs fixing.

輸入參數typelimitoffsetstatuspriorityproject_id必填

get_annotation

Get annotation

Fetch one annotation with its full captured context: page, target element (selector, XPath, styles), surrounding DOM, and runtime diagnostics (console + network).

輸入參數annotation_id必填

list_comments

List comments

Read the discussion thread on an annotation.

輸入參數annotation_id必填

get_annotation_analysis

Get AI diagnosis

Read the latest stored AI diagnosis for an annotation (root causes, confidence, suggested fix, and likely source files). Returns null when none has been generated yet; run diagnose_annotation to create one.

輸入參數annotation_id必填

get_annotation_screenshots

Get screenshots

Return the region screenshots attached to an annotation as viewable images, so you can see exactly what the reporter pointed at.

輸入參數annotation_id必填

list_annotation_issues

List annotation issues

List the tracker (GitHub) issues opened from an annotation, with each issue’s live state. Set refresh to reconcile against GitHub (open/closed + any linked pull request).

輸入參數refreshannotation_id必填

list_project_issues

List project issues

List every tracker issue across a project, with an open/closed/agent summary.

輸入參數project_id必填

generate_prompt

Generate coding prompt

Render a ready-to-paste coding prompt for an annotation, formatted for a target assistant. A deterministic template (no AI spend); for a deeper analysis use diagnose_annotation.

輸入參數targetannotation_id必填

get_metrics

Get metrics

Overview of your projects: KPI counts, status breakdown, a daily time series, and recent activity.

無需輸入參數。

list_members

List members

List a project’s team members, including role and status.

輸入參數project_id必填

get_activity

Get activity

Recent activity feed for a project (status changes, comments, issues, members).

輸入參數project_id必填

update_annotation_status

Update status

Update an annotation’s status, e.g. to mark it in_progress or completed.

輸入參數status必填annotation_id必填

add_comment

Add comment

Add a comment to an annotation’s thread, e.g. to record what you changed. Notifies the team.

輸入參數body必填annotation_id必填

diagnose_annotation

Run AI diagnosis

Run the AI Diagnosis Engine on an annotation: it analyses the captured page, element, DOM, and runtime errors to produce ranked root causes, a confidence score, a suggested fix, and the likely source files. Calls an AI provider (uses tokens / may cost money) and stores the result. Requires AI + diagnosis enabled by the workspace admin and a key set.

輸入參數modelproviderannotation_id必填

create_github_issue

Create GitHub issue

Create a GitHub issue from an annotation. Uses the project’s configured repo and assignee unless overridden. Embeds the generated AI prompt in the issue body.

輸入參數repoassigneesannotation_id必填

版本歷史

1
  1. v0.1.4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