Guardian MCP
Allows users to interact with Guardian through a Telegram bot, including checking account, portfolio, and policy details, previewing trades, and confirming orders before submission to Alpaca paper trading.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Guardian MCPPropose a paper trade to buy $200 worth of AAPL."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Guardian MCP
Guardian MCP is a policy-bounded AI trading gateway for Alpaca paper trading. AI agents can propose stock, ETF, and crypto trades, but deterministic rules decide whether the trade is blocked, approved for human review, or submitted to Alpaca paper trading.
Why this exists
Most AI trading demos give the model too much authority. Guardian separates the workflow:
AI proposes a TradeIntent
Guardian validates the intent
Policy engine approves or blocks
Alpaca paper trading receives only approved low-risk orders
Every decision becomes an audit receiptRelated MCP server: alpaca-mcp-server
Current scope
US equities and ETFs
Crypto pairs on Alpaca, including
BTC/USD,ETH/USD, andSOL/USDPaper trading only
Mock Alpaca execution when API keys are missing
Local JSON audit log for the first demo
Official Alpaca Skills alignment for paper-only safety gates
Getting started
npm install
npm run devOpen http://localhost:3000.
To connect Alpaca paper trading, create .env.local:
ALPACA_API_KEY=your_key
ALPACA_SECRET_KEY=your_secret
ALPACA_PAPER_BASE_URL=https://paper-api.alpaca.markets
ALPACA_PAPER_TRADE=true
TELEGRAM_BOT_TOKEN=your_botfather_tokenAlpaca's official skills also use APCA_API_KEY_ID and APCA_API_SECRET_KEY; Guardian supports those names too.
Do not paste API keys into chat or commit them. Guardian blocks live Alpaca configuration and only supports paper trading.
Demo flow
Pick a mock AI trade proposal.
Run the guard.
See allowed trades submitted to mock or real Alpaca paper trading.
See blocked trades recorded with exact policy reasons.
Telegram bot
Guardian can also run as a Telegram bot for the chat-first demo:
npm run telegramBot commands:
/account
/portfolio
/policy
/preview buy SOL/USD 50
/confirm guardian-tg-...The bot uses long polling, so it should run on your Google VM, not Vercel. It never submits silently: /preview creates the paper order preview and /confirm is required before Guardian can submit to Alpaca paper trading.
To run beside another process on the VM:
pm2 start npm --name guardian-telegram -- run telegram
pm2 savePolicy defaults
Paper-only mode
Max trade notional:
$500Human approval above:
$100Max daily loss:
2%Max position exposure:
15%Max crypto trade notional:
$100Max crypto portfolio exposure:
20%Crypto cooldown:
15 minutes
Hackathon pitch
Guardian MCP lets AI trading agents operate through Alpaca without giving the model unchecked control over an account. The model can research and propose trades; Guardian enforces deterministic user policy before execution and records every decision in a replayable audit trail.
Alpaca Skills
This repo is aligned with Alpaca's official agent skills. See docs/alpaca-skills.md for the paper-trading guardrails and the remaining Alpaca-specific work.
MCP Server
Guardian includes its own MCP server:
npm run mcpSee docs/mcp.md for tool names and MCP client configuration.
This server cannot be deployed
Maintenance
Related MCP Connectors
Automate trading on your own Alpaca account - build, backtest and run strategies via your AI.
Unified financial infrastructure connecting AI agents directly to trade live/demo brokerage accounts, Web3 non-custodial wallets, real-time market data across equities, ETFs, crypto, forex, options, DeFi swaps, and prediction markets, institutional research feeds, and algorithmic strategy backtesters.
- CPZAIOAuthcom.cpz-lab.mcp
Build, backtest, and deploy quantitative trading strategies from your AI agent.
Pre-execution governance for AI agents. Deterministic PASS/FAIL/REVIEW verdicts, replayable proof.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to execute stock trading operations with built-in risk controls and human approval workflows. Supports paper trading simulation, real brokerage integration (Alpaca, Tradier), backtesting, sentiment analysis, and portfolio management while maintaining strict separation between AI intelligence and trade execution.MIT
- AlicenseNot gradedqualityDmaintenanceEnables natural language trading operations for stocks, options, crypto, and portfolio management via Alpaca's Trading API through AI assistants.MIT
- AlicenseBqualityBmaintenanceProvides a safe trading toolkit for AI agents with market data, indicators, paper trading, and guarded broker integration.63MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to analyze Ethereum wallets, simulate transactions, and draft transfers with deterministic policy and risk scoring, requiring human approval before on-chain execution.7 npmISC