Tradewink

io.github.leancoderkavy由 GitHub 帳號 leancoderkavy 發布v0.3.0更新於 Oct 28, 2026

MCP for Tradewink trading signals and watchlists. Not financial advice.

已驗證Streamable HTTP可網頁執行金融資料與分析

概覽

AI 產生的概覽

Tradewink MCP 讓助理存取 Tradewink 的交易訊號、行情資料、自選清單、投資組合與下單工具。

可以做什麼
  • 讀取帳戶、持倉與即時投資組合資料
  • 管理自選清單：查看、新增與移除標的
  • 取得行情：報價、批次報價、歷史資料、新聞、分析師評等、財報與技術分析
  • 篩選市場異動與篩選器結果
  • 下單、平倉、查看未成交委託與取消委託
  • 查看策略、交易歷史、分析統計、成交紀錄與稽核日誌
適用情境
適合想讓助理在聊天用戶端中做市場研究、查看 Tradewink 投資組合、管理自選清單或依 Tradewink 訊號行動的情境。README 也說明可與獨立的券商伺服器搭配作為執行層。
執行需求
需要 Tradewink 帳戶與授權：首次執行透過 OAuth 2.1 瀏覽器授權，或在 TRADEWINK_API_KEY 提供手動 API 金鑰。README 中的本機套件以 npx（Node.js）透過 stdio 執行並代理至遠端端點；清單列出需要 Authorization bearer 權杖的遠端 streamable HTTP 端點。可用 TRADEWINK_ENDPOINT 或 --endpoint 覆寫伺服器位址。
安裝前請注意
此伺服器可下單、平倉與取消未成交委託，作用於與券商連結的帳戶，助理可能實際動用資金。OAuth 權杖會快取在本機磁碟並可自動更新 60 天；手動 API 金鑰提供長期存取權限。README 說明訊號僅供參考，可能不完整、延遲或錯誤，並強調須自行檢視每個訊號與委託。不構成投資建議。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Tradewink，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/io.github.leancoderkavy%2Ftradewink/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": {
    "tradewink": {
      "type": "http",
      "url": "https://tradewink-mcp.fly.dev/mcp",
      "headers": {
        "Authorization": "Bearer <YOUR_TOKEN>"
      }
    }
  }
}

README

@tradewink/mcp

Connect any LLM to Tradewink signals and analysis via the Model Context Protocol, then pair it with Alpaca for agent-driven execution.

Get real-time quotes, manage your portfolio, execute trades, run AI analysis, and more — all from your AI assistant.

Quick Start

1. Configure Your Client

No API key needed — @tradewink/mcp uses OAuth 2.1. The first time you run it, a browser window opens to authorize your tradewink.com account. Tokens are cached in ~/.tradewink/mcp/ and refresh automatically.

Automatic Setup (Recommended)

Auto-detect and configure all your installed AI clients with one command:

bash
npx @tradewink/mcp setup

This detects and configures: Claude Code, Claude Desktop, VS Code (Copilot), Cursor, Windsurf, and OpenAI Codex CLI.

Manual Setup
Claude Code
bash
claude mcp add tradewink -- npx -y @tradewink/mcp
Claude Desktop

Add to your claude_desktop_config.json:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
json
{  "mcpServers": {    "tradewink": {      "command": "npx",      "args": ["-y", "@tradewink/mcp"]    }  }}
VS Code (Copilot)

Add to your VS Code settings (.vscode/settings.json or global settings):

json
{  "mcp": {    "servers": {      "tradewink": {        "command": "npx",        "args": ["-y", "@tradewink/mcp"],        "env": {          "TRADEWINK_API_KEY": "tw_your_key_here"        }      }    }  }}
Cursor

Add to .cursor/mcp.json (project) or ~/.cursor/mcp.json (global):

json
{  "mcpServers": {    "tradewink": {      "command": "npx",      "args": ["-y", "@tradewink/mcp"]    }  }}
OpenAI Codex CLI
bash
codex mcp add tradewink -- npx -y @tradewink/mcp
Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

json
{  "mcpServers": {    "tradewink": {      "command": "npx",      "args": ["-y", "@tradewink/mcp"]    }  }}
Any MCP-Compatible Client

Any client that supports MCP stdio servers works:

bash
npx @tradewink/mcp

2. Start Using It

Ask your AI assistant things like:

  • "What's my account balance?"
  • "Show me my open positions"
  • "Get a quote for AAPL"
  • "What are today's top market movers?"
  • "Analyze NVDA with AI"
  • "Add TSLA to my watchlist"
  • "What are my active trading strategies?"
  • "Show my trade history for the past week"

Tradewink Signals + Alpaca Execution

Use tradewink-mcp as the signal and research layer for your agent, and the official Alpaca MCP server as the execution layer. This keeps Tradewink's market context, screeners, technical analysis, and AI insights separate from the broker tools that place orders.

This is an integration pattern: no Alpaca order-execution code is added to tradewink-mcp.

Add both servers to the same Claude Desktop or Cursor MCP configuration:

json
{  "mcpServers": {    "tradewink": {      "command": "npx",      "args": ["-y", "@tradewink/mcp"]    },    "alpaca": {      "command": "uvx",      "args": ["alpaca-mcp-server"],      "env": {        "ALPACA_API_KEY": "your_paper_api_key",        "ALPACA_SECRET_KEY": "your_paper_secret_key",        "ALPACA_PAPER_TRADE": "true"      }    }  }}

The official Alpaca server uses uvx and defaults to paper trading; the explicit ALPACA_PAPER_TRADE setting above makes the safety boundary visible. Create paper keys in Alpaca's dashboard and keep credentials in your local MCP client configuration. See the Alpaca MCP setup and configuration docs.

Example agent workflow

Ask the agent to:

  1. Get a Tradewink signal, such as market movers plus get_technical_analysis or analyze_ticker for AAPL.
  2. Check the current quote and paper-account buying power, define a maximum loss/risk budget, and size the position from that budget rather than blindly copying a signal.
  3. If the thesis and sizing pass review, place a paper order with Alpaca's place_stock_order, then report the order details and risk controls.

For example: “Use Tradewink to find a bullish AAPL setup, explain the signal and invalidation level, size it to no more than 1% of my paper account, and—only after I confirm—place the corresponding paper order through Alpaca.”

Tradewink signals are informational and may be incomplete, delayed, or wrong. Trading involves substantial risk, including loss of principal; an agent can misunderstand data or instructions. Review every signal, size, and order yourself, and do not switch ALPACA_PAPER_TRADE to false until you understand and accept the risks.

Available Tools

CategoryTools
Accountget_account, get_positions, get_portfolio_live
Watchlistget_watchlist, add_to_watchlist, remove_from_watchlist
Market Dataget_quote, batch_quote, get_historical, get_news, get_analyst_ratings, get_earnings, get_technical_analysis
Screenerget_market_movers, get_screener_results
Tradingplace_order, close_position, get_open_orders, cancel_order
Strategiesget_strategies, get_strategy, get_exit_strategies, get_sell_confidence
Trade Historyget_trades, get_trade_analytics, get_trade_executions, get_audit_log
AI Analysisask_ai, analyze_ticker, get_insights
Preferencesget_preferences, list_preference_options, update_preference, reset_preference
Schedulingschedule_analysis, list_scheduled_jobs, cancel_scheduled_job

Configuration

Environment VariableRequiredDefaultDescription
TRADEWINK_API_KEYNo—Skip the OAuth flow and use a manual API key (tw_...). Useful for headless CI.
TRADEWINK_ENDPOINTNohttps://tradewink-mcp.fly.dev/mcpCustom MCP server endpoint

You can also pass --endpoint as a CLI argument:

bash
npx @tradewink/mcp --endpoint https://custom.example.com/mcp

Auth

OAuth 2.1 is the default. On first run a browser window opens to tradewink.com/mcp/authorize, you approve the request, and the bridge caches an access + refresh token pair at ~/.tradewink/mcp/<host>.json. Access tokens live for 1 hour and refresh automatically for 60 days. You can revoke access anytime from your dashboard.

Set TRADEWINK_API_KEY=tw_... to skip OAuth and use a manual key (e.g. for CI, sandboxes, or pinned long-lived access).

How It Works

This package runs a local MCP server over stdio and proxies every tool/resource call to the remote Tradewink MCP server over HTTPS. OAuth tokens are minted per-device and scoped to your account.

Your LLM Client ←stdio→ @tradewink/mcp ←HTTPS (Bearer token)→ Tradewink MCP Server

License

MIT

來源：README.md，提交 3cf1420

工具

0
SourceWeft 尚未讀取此伺服器的工具。

版本歷史

1
  1. v0.3.0最新Sep 24, 2026

評分

已滿足適用檢查項的 85%

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

    請讓伺服器在無憑證時回應 tools/list，或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分