Profitelligence MCP Server
Financial intelligence for AI agents. Give Claude access to insider trading data, SEC filings, economic indicators, and multi-signal analysis — all through a single MCP server.
README
<p align="center"> <img src="assets/logo.png" alt="Profitelligence" width="200" /> </p>
<h1 align="center">Profitelligence MCP Server</h1>
<p align="center">
Financial intelligence for AI agents. Give Claude access to insider trading data, SEC filings, economic indicators, and multi-signal analysis — all through a single MCP server.
Get Started Free · View Pricing
Why Profitelligence?
Traditional financial APIs return mountains of raw data. Your AI agent burns through tokens parsing CSVs, making repeated calls, and piecing together context. Profitelligence is different.
We provide semantically dense, LLM-optimized responses designed for AI agents:
| Traditional MCP Servers | Profitelligence MCP |
|---|---|
| Many narrow tools (10-20+) | 7 powerful tools with rich filtering |
| Raw data dumps | Pre-contextualized intelligence |
| Multiple calls to answer one question | One call, complete answer |
| Token-heavy responses | Optimized for token efficiency |
| Complex tool orchestration | Agent has full control |
What You Get
- Insider Trading Intelligence — Form 4 filings with entity search ("find Warren Buffett's trades")
- SEC Filing Analysis — AI-summarized 8-K events with impact scoring
- Market Data — Company profiles, OHLC prices, sector/industry context
- Economic Indicators — Federal Reserve (FRED) data for macro context
- Multi-Signal Analysis — Opportunity scoring that combines insider activity + SEC events + price action (PRO)
Safe by Design
This MCP server is a thin, stateless layer over Profitelligence's REST APIs:
- ✅ Read-only — No ability to modify your account or execute trades
- ✅ No data storage — Responses flow through, nothing is cached locally
- ✅ Open source — Audit the code yourself, it's ~2000 lines of Python
- ✅ Your API key — You control access; revoke anytime from your dashboard
Quick Start
Option A: Connect via MCP Directory (Recommended)
The easiest way to get started:
- Visit the Anthropic MCP Directory
- Find and click Profitelligence
- Complete OAuth authentication with your Google account
- Start asking financial questions!
No API key needed — authentication is handled automatically.
Option B: Manual Setup with API Key
1. Create a Free Account
Sign up at profitelligence.com — no credit card required.
The free tier includes:
- Access to top 500 stocks by volume
- Insider trading data (Form 4)
- SEC 8-K filing summaries
- Economic indicators (FRED)
- 250 API calls/day
2. Get Your API Key
Generate an API key from your account dashboard.
3. Add to Claude
Claude Desktop (easiest):
Go to Settings → Connectors → Add Custom Connector and paste this URL:
https://mcp.profitelligence.com/mcp?apiKey=YOUR_API_KEY
Replace YOUR_API_KEY with your key (e.g. pk_live_abc123...). That's it — start chatting.
Claude Desktop (config file alternative):
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"profitelligence": {
"url": "https://mcp.profitelligence.com/mcp?apiKey=YOUR_API_KEY",
"transport": "streamable-http"
}
}
}
Claude Code (one-liner):
claude mcp add profitelligence --transport http \
https://mcp.profitelligence.com/mcp \
--header "X-API-Key: YOUR_API_KEY"
That's it! Ask Claude: "What insider buying happened in tech stocks this week?"
Pricing & Tiers
| Feature | Free | PRO ($4.99/mo) | Elite ($9.99/mo) |
|---|---|---|---|
| Stock coverage | Top 500 | Top 6,000 | All stocks |
| API calls | 250/day | 1,000/day | 10,000/day |
| Insider trading (Form 4) | ✅ | ✅ | ✅ |
| SEC 8-K summaries | ✅ | ✅ | ✅ |
| Economic indicators | ✅ | ✅ | ✅ |
| Opportunity scoring | — | ✅ | ✅ |
| Multi-signal analysis | — | ✅ | ✅ |
| Alert subscriptions | — | 10 alerts | Unlimited |
| Priority support | — | — | ✅ |
Early adopter pricing — $4.99/mo PRO and $9.99/mo Elite rates are locked in for the first 1,000 users (regular: $10/mo and $29/mo). Upgrade anytime.
Available Tools
7 tools designed for maximum token efficiency. Each answers complete questions with rich, contextual responses:
| Tool | Purpose | Example Query |
|---|---|---|
pulse() |
Market snapshot | "What's happening in the market right now?" |
investigate(subject) |
Deep research | "Tell me about NVDA's insider activity" |
screen(focus) |
Opportunity scanning | "Find tech stocks with insider buying" |
assess(symbol) |
Position health check | "How is my AAPL position doing?" |
institutional(query_type) |
13F intelligence | "What is Berkshire buying?" |
search(q) |
Semantic search | "Find CEO resignation filings" |
service_info() |
Account & service info | "What's my subscription tier?" |
Usage Examples
Example 1: Get Market Overview
User prompt: "What's happening in the market right now?"
What happens:
- Server calls
pulse()tool - Returns current market movers (gainers, losers, volume leaders)
- Shows recent high-impact SEC 8-K filings with AI-generated summaries
- Lists notable insider transactions from Form 4 filings
- Provides key economic indicators from FRED
Example 2: Research a Company
User prompt: "Tell me about NVIDIA's insider activity and recent filings"
What happens:
- Server calls
investigate("NVDA")tool - Returns company profile with sector, market cap, and CEO
- Shows price history and technical signals
- Lists recent 8-K material events with impact scores
- Provides Form 4 insider transaction summary and sentiment
- Includes latest financials from 10-K/10-Q filings
Example 3: Find Investment Opportunities
User prompt: "Find tech stocks where insiders are buying"
What happens:
- Server calls
screen(focus="insider", sector="Technology")tool - Scans market for insider buying clusters in Technology sector
- Returns ranked list of opportunities with transaction details
- Shows insider names, roles, and transaction values
- Provides opportunity scores based on multi-signal analysis
Example 4: Track Institutional Investors
User prompt: "What is Berkshire Hathaway buying?"
What happens:
- Server calls
institutional("manager", identifier="Berkshire")tool - Uses fuzzy search to find matching institutional manager
- Returns top holdings from latest 13F filing
- Shows recent position changes (new buys, additions, exits)
- Lists high-conviction positions (5%+ of portfolio)
Tool Reference
Understanding when to use each tool is key to getting the best results. Here's the mental model:
pulse() — What's happening right now?
Use when: User wants a market overview without a specific company in mind.
Returns: Market movers, recent high-impact 8-K filings, notable insider trades, key economic indicators.
No parameters. Just call it.
pulse()
Best for: Starting a session, getting oriented, "what should I look at today?"
investigate(subject) — Deep dive on a specific entity
Use when: User names a specific company, insider, or sector they want to research.
Auto-detects entity type:
- Stock symbols (
AAPL,NVDA) → Company research - CIK numbers (
0001067983) → Insider research - Sector names (
Technology) → Sector analysis
Returns for companies:
- Profile, price history, material events (8-K filings)
- Insider intelligence (Form 4 summary + recent transactions)
- Financials (income statement, balance sheet, cash flow from 10-K/10-Q)
- Opportunity signals, technical indicators
investigate("AAPL") # Company
investigate("0001067983") # Warren Buffett by CIK
investigate("Technology", entity_type="sector") # Sector
Best for: "Tell me about X", "What's going on with X?", research before a decision.
screen(focus) — Find opportunities across the market
Use when: User wants to discover stocks matching certain criteria, not research a specific one.
Parameters:
focus: What signal to screen for"all"— Everything (default)"multi_signal"— Stocks with multiple confirming signals"insider"— Insider buying clusters"events"— High-impact 8-K filings
sector: Filter to a sector (e.g.,"Technology")min_score: Minimum opportunity score (0-100)days: Lookback period (default 7)limit: Max results (default 25)
screen() # Everything
screen(focus="insider", sector="Technology") # Tech insider buying
screen(focus="events", days=14) # Recent material events
Best for: "What opportunities are out there?", "Find me stocks where insiders are buying."
assess(symbol) — Health check for a position you own
Use when: User owns a stock and wants to evaluate whether to hold, add, or sell.
Returns:
- Current price action and technical signals
- Recent material events (8-K filings)
- Insider sentiment (are insiders buying or selling?)
- Institutional sentiment (13F accumulation/distribution)
- Financials snapshot
- Risk factors
assess("NVDA")
assess("AAPL", days=90) # Longer lookback
Best for: Portfolio review, "How's my position in X?", "Should I be worried about X?"
institutional(query_type) — 13F institutional investor intelligence
Use when: User wants to know what big money is doing.
Query types:
"manager"— Profile a specific fund (by name or CIK)"security"— Who owns a specific stock?"signal"— Find stocks with institutional flow patterns
Signal types (for query_type="signal"):
"accumulation"— Stocks institutions are buying"distribution"— Stocks institutions are selling"conviction"— High conviction positions (5%+ of portfolio)"new"— Fresh institutional positions
institutional("manager", identifier="Citadel") # What does Citadel own?
institutional("security", identifier="NVDA") # Who owns NVIDIA?
institutional("signal", signal_type="accumulation") # Smart money buying
Best for: "What is Berkshire buying?", "Who owns NVDA?", "Follow the smart money."
search(q) — Find anything across the platform
Use when: User is looking for something specific but doesn't know where it is.
Searches across:
- SEC 8-K filings (with LLM-extracted insights)
- Companies (by name or symbol)
- Insiders (executives, directors)
- Institutional managers (13F filers)
Parameters:
q: Search query (min 2 chars)entity_type: Filter results —"filing","company","insider","manager"sector: Filter by sectorimpact: Filter filings by impact level —"HIGH","MEDIUM","LOW"limit: Max results (default 20, max 100)
search("CEO resignation") # Find CEO departure filings
search("NVIDIA", entity_type="company") # Find NVIDIA
search("Buffett", entity_type="insider") # Find Warren Buffett
search("activist", entity_type="manager") # Find activist funds
search("acquisition", sector="Technology", impact="HIGH") # High-impact tech M&A
Best for: "Find filings about X", "Search for Y", discovery when you're not sure where to look.
service_info(info_type) — Account and service information
Use when: User asks about their subscription, features, or the service itself.
Info types:
"overview"— Service description and capabilities (default)"profile"— User's subscription tier, features, account status"pricing"— Subscription tiers and pricing"capabilities"— Available tools and data sources"status"— Server health and configuration
service_info() # Overview
service_info("profile") # Your account
service_info("pricing") # Pricing info
Best for: "What can you do?", "What's my tier?", "How much does PRO cost?"
Tool Selection Guide
| User Intent | Tool |
|---|---|
| "What's happening?" / "Market overview" | pulse() |
| "Tell me about [company/insider/sector]" | investigate() |
| "Find stocks where..." / "What opportunities..." | screen() |
| "How's my [position]?" / "Should I worry about..." | assess() |
| "What is [fund] buying?" / "Who owns [stock]?" | institutional() |
| "Search for..." / "Find filings about..." | search() |
| "What's my tier?" / "What can you do?" | service_info() |
MCP Prompts
Pre-built research workflows that produce complete reports:
insider_activity_report— Insider trading analysis for a symbolsec_filing_intelligence_report— 8-K filing analysismacro_market_conditions_report— Economic backdropinfluential_investor_research— Smart money deep divesector_opportunity_scan— Multi-signal sector analysisquarterly_stock_checkup— Portfolio position review
Self-Hosting (Optional)
Most users don't need this. The hosted version at
mcp.profitelligence.comis free, fast, and requires zero setup. Just add your API key and go (see Quick Start above).
This repo exists so you can:
- Audit the code — See exactly what runs when you use our MCP server
- Self-host for privacy — Run your own instance if you prefer
- Customize — Fork and add your own tools
When to Self-Host
| Use Case | Recommendation |
|---|---|
| Just want financial intelligence in Claude | Use hosted — it's free |
| Want to audit what code is running | Read this repo, use hosted |
| Need custom web search privacy | Self-host with your own Brave API key |
| Building on top of our tools | Fork and extend |
Docker Hub
We publish multi-arch images (amd64 + arm64) to Docker Hub:
docker pull profitelligence/mcp-server:latest
Run in stdio mode (for Claude Desktop):
docker run -i --rm \
-e PROF_API_KEY=pk_live_xxx \
profitelligence/mcp-server:latest
Run in HTTP mode (for development/testing):
docker run -d -p 3000:3000 \
-e PROF_API_KEY=pk_live_xxx \
-e PROF_MCP_MODE=http \
profitelligence/mcp-server:latest
From Source
git clone https://github.com/profitelligence/profitelligence-mcp-server.git
cd profitelligence-mcp-server
pip install -e .
PROF_API_KEY=pk_live_xxx python -m src.server
Environment Variables
| Variable | Required | Description |
|---|---|---|
PROF_API_KEY |
Yes | Your API key from profitelligence.com |
PROF_MCP_MODE |
No | stdio (default) or http |
PROF_MCP_PORT |
No | HTTP port (default: 3000) |
BRAVE_API_KEY |
No | Your Brave Search key for web search privacy |
Web Search Privacy
The hosted version uses our shared infrastructure. When self-hosting, you can provide your own Brave Search API key (2,000 free queries/month):
- ✅ Your search queries go directly to Brave
- ✅ We never see what you're searching for
- ✅ Falls back to DuckDuckGo if no key provided
See CONTRIBUTING.md for development setup.
Privacy Policy
See our privacy policy: profitelligence.com/privacy
Support
- Issues: GitHub Issues
- Contact: profitelligence.com/contact
- Email: support@profitelligence.com
License
MIT License - See LICENSE for details.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。