Systemr Python MCP Server

systemr

Build AI trading agents with institutional-grade risk management.

Install

pip install systemr

Requires Python 3.9 or higher. The only dependency is httpx.

Your first trade gate

from systemr import SystemRClient

client = SystemRClient(api_key="sr_agent_...")

gate = client.pre_trade_gate(
    symbol="AAPL",
    direction="long",
    entry_price="185.50",
    stop_price="180.00",
    equity="100000",
)

if gate["gate_passed"]:
    print(f"Buy {gate['sizing']['shares']} shares of AAPL")
    print(f"Risk: ${gate['sizing']['risk_amount']}")

Five lines of code. Position sizing, risk validation, and system health in a single call -- free.

Chat

Talk to your agent in natural language. The LLM interprets your intent and calls the right tools.

from systemr import SystemRClient

client = SystemRClient(api_key="sr_agent_...")

# Ask a question — the agent picks the right tools
resp = client.chat("What's my current G-score? I've had these R-multiples: 1.5, -1.0, 2.0, -0.5, 1.8")
print(resp["text"])
print(f"Credits used: ${resp['credits_used']}")

# Continue the conversation
resp = client.chat("Run a Monte Carlo on those same trades", session_id=resp.get("session_id"))
print(resp["text"])

What you get

68 free tools across 13 categories. 25 brokers and exchanges. All tools are free. You only pay when an LLM is used. Pay via Stripe, SOL, USDC, USDT, PYUSD, or OSR (50% bonus).

Connect to 25 brokers including IBKR, Schwab, Binance, Coinbase, Alpaca, Kraken, dYdX, Polymarket, and more.

---

Tools

Every tool is callable via client.call_tool("tool_name", **kwargs).

Core (4 tools)

Analysis (18 tools)

Intelligence (11 tools)

Planning (4 tools)

Data (3 tools)

System (5 tools)

Compound (2 tools)

Journal (1 tool)

Memory and ML (7 tools)

---

Brokers and Exchanges

25 brokers and exchanges across every major market.

Traditional Brokers

Forex

Crypto Exchanges

DeFi Protocols

Prediction Markets

---

Broker Connection Example

import requests

headers = {"Authorization": "Bearer sr_agent_..."}
base = "https://agents.systemr.ai"

# Connect to Binance
resp = requests.post(f"{base}/v1/broker/connect", headers=headers, json={
    "broker": "binance",
    "credentials": {
        "api_key": "your_binance_api_key",
        "api_secret": "your_binance_api_secret",
    },
})
print(resp.json())  # {"status": "connected", "broker": "binance"}

# Check positions
resp = requests.get(f"{base}/v1/broker/positions", headers=headers)
positions = resp.json()
for pos in positions["positions"]:
    print(f"{pos['symbol']}: {pos['quantity']} @ {pos['avg_price']}")

# Place an order
resp = requests.post(f"{base}/v1/broker/order", headers=headers, json={
    "symbol": "BTCUSDT",
    "side": "BUY",
    "type": "LIMIT",
    "quantity": "0.01",
    "price": "65000.00",
})
print(resp.json())  # {"order_id": "...", "status": "NEW"}

---

Workflows

1. Pre-Trade Gate (call before every trade)

One call, three checks: position sizing, risk validation, and system health.

from systemr import SystemRClient

client = SystemRClient(api_key="sr_agent_...")

gate = client.pre_trade_gate(
    symbol="AAPL",
    direction="long",
    entry_price="185.50",
    stop_price="180.00",
    equity="100000",
    r_multiples=["1.5", "-1.0", "2.0", "-0.5", "1.8", "0.8", "-0.3"],
)

if gate["gate_passed"]:
    shares = gate["sizing"]["shares"]
    risk = gate["sizing"]["risk_amount"]
    print(f"APPROVED: Buy {shares} shares, risking ${risk}")
    if gate.get("system_health"):
        print(f"System G: {gate['system_health']['g']}")

Cost: Free

2. Backtest Diagnostic (6-tool chain)

Feed your backtest R-multiples and get a complete system analysis: equity curve, grade, drawdown, Monte Carlo projection, variance killers, and win/loss distribution.

diag = client.run_backtest_diagnostic(
    r_multiples=[
        "1.5", "-1.0", "2.0", "-0.5", "1.8",
        "0.8", "-0.3", "2.5", "-1.0", "1.2",
        "0.5", "-0.8", "1.1", "-0.2", "3.0",
    ],
    starting_equity="100000",
)

print(f"Grade: {diag['system_r_score']['grade']}")
print(f"Total return: {diag['equity_curve']['total_return']}")
print(f"Max drawdown: {diag['equity_curve']['max_drawdown_pct']}")
print(f"Monte Carlo median: ${diag['monte_carlo']['median_final_equity']}")
print(f"Variance killers: {len(diag['variance_killers'].get('killers', []))} trades")
print(f"Win rate: {diag['win_loss']['win_rate']}")

Cost: Free

3. Full System Assessment

Quarterly system review. Runs G-metrics, win/loss analysis, Kelly criterion, Monte Carlo, drawdown, and what-if scenarios in a single compound call.

assessment = client.assess_system(
    r_multiples=[
        "1.5", "-1.0", "2.0", "-0.5", "1.8",
        "0.8", "-0.3", "2.5", "-1.0", "1.2",
        "0.5", "-0.8", "1.1", "-0.2", "3.0",
        "-1.0", "0.7", "1.5", "-0.6", "2.1",
    ],
    starting_equity="100000",
)

print(f"Verdict: {assessment['verdict']}")
# STRONG_SYSTEM, VIABLE_SYSTEM, MARGINAL_SYSTEM, or FAILING_SYSTEM

print(f"G metric: {assessment['g_metrics']['g']}")
print(f"Kelly fraction: {assessment['kelly']['kelly_fraction']}")
print(f"Half-Kelly: {assessment['kelly']['half_kelly']}")
print(f"Monte Carlo median: ${assessment['monte_carlo']['median_final_equity']}")
print(f"Max drawdown: {assessment['drawdown']['max_drawdown_pct']}")

Cost: Free

4. Market Intelligence Scan

Scan a watchlist for technical conditions, then score each matching signal for quality and confidence.

scan = client.run_market_scan(
    symbols=["AAPL", "MSFT", "GOOGL", "NVDA", "TSLA"],
    conditions=["rsi_oversold", "volume_spike"],
    market_data={
        "AAPL": {
            "indicators": {"rsi_14": "28", "relative_volume": "2.1"},
            "current_price": "180.00", "regime": "RANGING", "atr": "3.50",
        },
        "MSFT": {
            "indicators": {"rsi_14": "55", "relative_volume": "0.8"},
            "current_price": "400.00", "regime": "TRENDING_UP", "atr": "5.00",
        },
        "GOOGL": {
            "indicators": {"rsi_14": "22", "relative_volume": "1.8"},
            "current_price": "155.00", "regime": "TRENDING_DOWN", "atr": "4.20",
        },
    },
)

for signal in scan["scored_signals"]:
    print(f"{signal['symbol']}: quality={signal['quality_score']}/100, "
          f"confidence={signal['signal_confidence']}")

Cost: Free

5. Memory-Assisted Trading

Agents can persist memories across sessions and use behavioral analytics to improve over time.

# Store a trading insight
client.call_tool("store_memory",
    key="aapl_earnings_pattern",
    value="AAPL tends to gap down after earnings then recover within 3 days. "
          "Best entry is day 2 close if RSI < 35.",
)

# Later: recall it before an AAPL earnings trade
memories = client.call_tool("search_memory", query="aapl earnings")
print(memories)

# Check your behavioral biases
biases = client.call_tool("get_trading_biases",
    r_multiples=["1.5", "-1.0", "2.0", "-0.5", "0.3", "-2.0", "0.8"],
    trade_durations=["120", "30", "240", "15", "60", "180", "90"],
)
print(f"Detected biases: {biases['biases']}")

# Get your behavioral fingerprint
fingerprint = client.call_tool("get_behavioral_fingerprint",
    r_multiples=["1.5", "-1.0", "2.0", "-0.5", "1.8"],
    trade_durations=["120", "30", "240", "15", "60"],
    position_sizes=["100", "50", "200", "75", "150"],
)
print(f"Risk appetite: {fingerprint['risk_appetite']}")
print(f"Adaptability: {fingerprint['adaptability']}")

Cost: Free

---

Billing

All 68 tools are free. You only pay when an LLM is used. Cost varies by query complexity.

Agents start free. No signup credit needed — all 68 tools cost nothing. You only need a balance for LLM usage.

Billing is at the owner level: one balance shared across all your API keys. You can set per-key spending limits (daily/monthly caps, operation restrictions).

Payment methods

Crypto deposits require a tx_signature (Solana transaction signature) and amount.

Wallet linking

import requests

headers = {"Authorization": "Bearer sr_agent_..."}
base = "https://agents.systemr.ai"

# Link your Solana wallet
requests.post(f"{base}/v1/agents/link-wallet", headers=headers, json={
    "wallet_address": "YourSolanaPublicKey...",
})

# Deposit OSR tokens
requests.post(f"{base}/v1/billing/deposit-osr", headers=headers, json={
    "tx_signature": "your_solana_tx_signature",
    "amount": "10000",
})

# Check balance
resp = requests.get(f"{base}/v1/billing/balance", headers=headers)
print(resp.json())  # {"balance": "...", "currency": "USDC"}

OSR presale discount

Agents whose linked wallet participated in the OSR token presale receive a permanent 20% discount on every tool call. The discount is verified on-chain through the linked Solana wallet. The wallet is the identity.

---

API Reference

All endpoints are prefixed with https://agents.systemr.ai/v1.

Agent Management

Billing

Broker

Tools

Support

Guardian

Trust

Leaderboard

Graduation

Audit

Onboarding

---

MCP

System R is available as an MCP (Model Context Protocol) server for direct integration with LLM clients like Cursor, Windsurf, and other MCP-compatible tools.

Remote server (recommended)

Add this to your MCP client configuration:

{
  "mcpServers": {
    "systemr": {
      "transport": "sse",
      "url": "https://agents.systemr.ai/mcp/sse",
      "headers": {
        "X-API-Key": "sr_agent_..."
      }
    }
  }
}

Local stdio server

Run a local proxy:

pip install systemr "mcp>=1.8.0"
SYSTEMR_API_KEY=sr_agent_... python -m systemr.mcp_server

Configuration for local stdio mode:

{
  "mcpServers": {
    "systemr": {
      "command": "python",
      "args": ["-m", "systemr.mcp_server"],
      "env": {
        "SYSTEMR_API_KEY": "sr_agent_..."
      }
    }
  }
}

Streamable HTTP

For clients that support the streamable HTTP transport:

{
  "mcpServers": {
    "systemr": {
      "transport": "streamable-http",
      "url": "https://agents.systemr.ai/mcp/sse",
      "headers": {
        "X-API-Key": "sr_agent_..."
      }
    }
  }
}

All 68 free tools are available through MCP with full input schemas and descriptions.

---

Links

---

License

MIT License. Copyright (c) 2026 System R AI. See LICENSE for details.