cpzai-mcp-server
MCP server for CPZAI platform, enabling AI agents to manage trading strategies, run backtests, route orders across brokers, and access portfolios, risk analytics, and market data through natural language.
README
<h1 align="center">CPZAI MCP Server</h1>
<p align="center"> <strong>Model Context Protocol server for AI agent access to <a href="https://ai.cpz-lab.com">CPZAI</a></strong> </p>
<p align="center"> <a href="https://github.com/CPZ-Lab/cpzai-mcp-server"><img src="https://img.shields.io/badge/language-TypeScript-blue.svg" alt="TypeScript"></a> <a href="https://github.com/CPZ-Lab/cpzai-mcp-server/actions/workflows/ci.yml"><img src="https://github.com/CPZ-Lab/cpzai-mcp-server/actions/workflows/ci.yml/badge.svg" alt="CI"></a> <a href="https://modelcontextprotocol.io/"><img src="https://img.shields.io/badge/MCP-v1.0-green.svg" alt="MCP v1.0"></a> <img src="https://img.shields.io/badge/transport-Streamable_HTTP-orange.svg" alt="Streamable HTTP"> <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-blue.svg" alt="License: MIT"></a> </p>
A production-grade Model Context Protocol (MCP) server that exposes CPZAI capabilities — strategy management, backtesting, multi-broker order routing (Alpaca / Interactive Brokers / FIX), portfolios, risk analytics, market data — as tools for any MCP-compatible AI agent (Claude, Cursor, GPT, etc.).
The hosted endpoint at https://mcp.cpz-lab.com/mcp is the only supported way to connect. This repo exists for transparency: anyone wiring their broker API keys through an AI agent should be able to read the exact code that handles those keys.
Quick Start
You'll need a CPZ API key + secret. Create one at ai.cpz-lab.com/settings/api-keys.
Cursor
Add to .cursor/mcp.json:
{
"mcpServers": {
"cpzai": {
"url": "https://mcp.cpz-lab.com/mcp",
"headers": {
"X-CPZ-Key": "your_cpz_key",
"X-CPZ-Secret": "your_cpz_secret"
}
}
}
}
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"cpzai": {
"url": "https://mcp.cpz-lab.com/mcp",
"headers": {
"X-CPZ-Key": "your_cpz_key",
"X-CPZ-Secret": "your_cpz_secret"
}
}
}
}
Claude (web / mobile) via OAuth
The server speaks OAuth 2.1 + PKCE at https://mcp.cpz-lab.com/.well-known/oauth-authorization-server. Use Claude's "Connect a server" flow and point it at https://mcp.cpz-lab.com/mcp.
Architecture
Agent (Cursor / Claude / GPT / custom)
│
│ Streamable HTTP + X-CPZ-Key / X-CPZ-Secret
▼
mcp.cpz-lab.com
│
▼
Node.js MCP server (stateless) ← this repo
│
│ per-request API key validation
▼
CPZ Platform REST API
Design decisions:
- Stateless transport. Each
POST /mcpcreates a fresh server instance — no session tracking, scales horizontally without coordination. - Thin protocol adapter. Validates the API key, then proxies tool calls to the CPZ REST API. Zero business logic in the MCP layer; what you see in
src/is what runs in production. - Key auth. Agents authenticate with CPZ platform API keys (
X-CPZ-Key/X-CPZ-Secret) — the same keys used by the REST API and thecpzPython SDK. - Paper-only by default. Live trading on a strategy requires explicit per-strategy promotion in the platform.
Available tools
18 tools exposed today, organized by domain.
| Tool | Description | Read-only |
|---|---|---|
| Strategies | ||
list_strategies |
List trading strategies with filtering | ✓ |
get_strategy |
Get a specific strategy by ID | ✓ |
create_strategy |
Create a new trading strategy | |
update_strategy |
Update an existing strategy | |
| Backtests | ||
get_backtest_results |
List backtest run results | ✓ |
| Orders & trading | ||
list_orders |
List trading orders with filtering | ✓ |
place_order |
Place a new trading order | |
list_positions |
List current portfolio positions | ✓ |
sync_portfolio |
Trigger portfolio sync across brokers | |
list_accounts |
List connected trading accounts | ✓ |
| Market data | ||
get_market_data |
Real-time quotes (price, bid/ask, volume) | ✓ |
| Risk | ||
compute_risk |
Compute fresh risk snapshot (VaR, Sharpe, drawdown) | |
list_risk_snapshots |
List historical risk snapshots | ✓ |
| Execution | ||
execute_strategy |
Execute a strategy on the Python backend | |
| Webhooks | ||
list_webhooks |
List webhook subscriptions | ✓ |
create_webhook |
Subscribe to platform events | |
delete_webhook |
Remove a webhook subscription | |
| User | ||
get_profile |
Get authenticated user profile | ✓ |
Local development
npm install
export CPZ_API_BASE_URL=https://api.cpz-lab.com
export CPZ_SERVICE_KEY=your_service_key
npm run dev
# server starts on http://localhost:3001
Tests + type check
npm test # single run
npm run test:watch # watch mode
npx tsc --noEmit # type check
Build
npm run build # compiles TS → dist/
npm start # runs the compiled server
Docker
docker build -t cpzai-mcp-server .
docker run --rm -p 3001:3001 \
-e CPZ_API_BASE_URL=https://api.cpz-lab.com \
-e CPZ_SERVICE_KEY=your_service_key \
cpzai-mcp-server
Security model
- All connections HTTPS with TLS 1.3 at the edge.
- API credentials are forwarded in HTTP headers and never logged.
- Per-request API key validation against the platform's
api_keystable. - All data access is user-scoped via the
user_idderived from the validated key. - OAuth tokens (when used) are held in server memory only — never persisted.
See the privacy policy for the full data-handling story.
Self-hosting
We don't currently support self-hosting. Use the hosted endpoint at https://mcp.cpz-lab.com/mcp with your CPZ API key. The hosted server is the same code in this repo plus AWS infrastructure (ALB, ECS Fargate, WAF, secrets) that's not part of this repo.
If you have a strong reason to self-host (e.g. air-gapped trading desk), open an issue and we'll talk.
Contributing
Issues and PRs welcome. Before submitting:
npx tsc --noEmitmust passnpm testmust pass- Keep changes minimal and aligned with existing patterns
License
MIT © CPZ Capital Ltd.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.