Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
作者提供的说明(英文原文)
Stock verdicts, trend stage, P&F, breadth, macro regime, SEC filings. Free key: api.vanoe.ai/signup
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。
VANOE_API_KEY, VANOE_API_BASE_URL标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
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": {
"vanoe-intelligence": {
"command": "uvx",
"args": [
"vanoe-intelligence-mcp"
],
"env": {
"VANOE_API_KEY": "your-value",
"VANOE_API_BASE_URL": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add vanoe-intelligence --env VANOE_API_KEY=your-value --env VANOE_API_BASE_URL=your-value -- uvx vanoe-intelligence-mcp运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"vanoe-intelligence": {
"command": "uvx",
"args": [
"vanoe-intelligence-mcp"
],
"env": {
"VANOE_API_KEY": "your-value",
"VANOE_API_BASE_URL": "your-value"
}
}
}
}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": {
"vanoe-intelligence": {
"type": "stdio",
"command": "uvx",
"args": [
"vanoe-intelligence-mcp"
],
"env": {
"VANOE_API_KEY": "your-value",
"VANOE_API_BASE_URL": "your-value"
}
}
}
}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。
VANOE_API_BASE_URL env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 vanoe-ai 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。
你需要填写:VANOE_API_KEY
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2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 Python (uv)。
同一分类中有:Spec Driven Development, Remote Desktop Commander, SwarmMemo
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
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Talk with agent swarms: rooms, private messages and paid tasks for AI agents
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。
Get a free API key in ten seconds: [https://api.vanoe.ai/signup?src=mcp-readme](https://api.vanoe.ai/signup?src=mcp-readme) — 1,000 credits a month, no card. The key page also prints the hosted MCP URL to paste into Claude.ai or ChatGPT as a connector.
Exposes the Vanoe Market Intelligence API as tools any MCP-capable AI assistant (Claude Desktop, Claude Code, Cursor, etc.) can call directly: a composite 0–100 verdict per ticker, trend-stage analysis (current and a backtest-safe weekly history), Point & Figure signals, sector breadth, the macro risk regime, SEC insider and 8-K filings, the macro calendar, and your usage.
This is a thin authenticated client, not a free tier: every tool call is a real, metered call against your plan using your own API key, exactly like calling the REST API yourself. Get a free key (1,000 credits a month, no card) here: https://api.vanoe.ai/signup?src=mcp-readme
Informational data only. Not financial advice. See https://api.vanoe.ai/terms.
The same tools are served over MCP's streamable HTTP transport by the API itself, for clients that take a URL (Claude.ai custom connectors, ChatGPT, hosted agent platforms):
https://api.vanoe.ai/mcp with Authorization: Bearer , orhttps://api.vanoe.ai/mcp/k/ for clients that cannot send headers (the key page after signup prints this URL; treat it like a password).Stateless, JSON responses, metered exactly like the REST API. Use the package below when you want the server to run on your own machine.
Download vanoe-market-intelligence.mcpb and open it with Claude Desktop (macOS or Windows). It asks for your API key once, stores it securely, and installs the server itself using the uv runtime. Source of the bundle is in mcpb/; rebuild with npx @anthropic-ai/mcpb pack mcpb vanoe-market-intelligence.mcpb.
No install step is needed with uv — uvx fetches and runs the package on demand:
uvx vanoe-intelligence-mcp
Or install it permanently with pipx install vanoe-intelligence-mcp or pip install vanoe-intelligence-mcp, which puts a vanoe-intelligence-mcp command on your PATH.
Set one environment variable, VANOE_API_KEY. VANOE_API_BASE_URL is optional and defaults to https://api.vanoe.ai.
Claude Desktop (claude_desktop_config.json) and most other clients:
{
"mcpServers": { "vanoe": { "command": "uvx", "args": ["vanoe-intelligence-mcp"], "env": { "VANOE_API_KEY": "sk_live_your_key_here" } } } }
Claude Code:
claude mcp add vanoe -e VANOE_API_KEY=sk_live_your_key_here -- uvx vanoe-intelligence-mcp
Then ask your assistant things like *"What's the verdict on NVDA and why?"*, *"Which stage are AAPL, MSFT and AMD in?"*, or *"Is the macro regime risk-on right now?"* — it will call the tools below.
| Tool | API call | Credits |
|---|---|---|
get_verdict(ticker) | GET /v1/verdict/{ticker} | 5 |
get_stage(tickers) | GET /v1/signals/stage | 1 per ticker |
get_stage_history(ticker, weeks) | GET /v1/signals/stage/history | 5 |
get_pnf(tickers) | GET /v1/signals/pnf | 1 per ticker |
get_sector_breadth() | GET /v1/sector-breadth | 1 |
get_macro() | GET /v1/macro | 1 |
get_filings(ticker) | GET /v1/filings/{ticker} | 1 |
get_short_pressure(tickers) | GET /v1/signals/short-pressure | 1 per ticker |
get_calendar() | GET /v1/calendar | 1 |
get_usage() | GET /v1/usage | 0 |
Tools never raise. Any failure — missing key, network error, quota exceeded, an unknown ticker — comes back as {"error": {"code", "message", "http_status"}} carrying the API's own message, so the assistant can explain what happened instead of crashing the call.
Webhook management is deliberately not exposed: it is account administration, not something to do mid-conversation. Use the REST API for that.
git clone https://github.com/vanoe-ai/vanoe-intelligence-mcp.git
cd vanoe-intelligence-mcp python -m venv .venv && . .venv/bin/activate # Windows: .venv\Scripts\activate pip install -e '.[test]' pytest tests VANOE_API_BASE_URL=http://localhost:8000 VANOE_API_KEY=sk_live_… vanoe-intelligence-mcp
Source lives at https://github.com/vanoe-ai/vanoe-intelligence-mcp; issues and pull requests welcome. If you also run the API server locally, install this in its own virtualenv, separate from the API server's own requirements.txt: the mcp SDK pulls in a newer starlette than the API's FastAPI pin tolerates, and sharing one environment breaks the API server with Router.__init__() got an unexpected keyword argument 'on_startup'. End users never hit this (the MCP server runs on their machine, next to their assistant).
tests/manual_e2e_smoke.py runs the real server as a subprocess over stdio against a reachable API with a real key; it is not part of pytest.
The MCP Python SDK is on v2: the server class is mcp.server.mcpserver.MCPServer (older tutorials show FastMCP at a different import path).
python -m build
twine upload dist/*
Bump version in pyproject.toml first. The older INTEL_API_KEY / INTEL_API_BASE_URL variable names are still accepted for compatibility.