System R AI

Risk management, planning, execution, data, analysis with persistent memory and encrypted privacy through one API.

55 tools. 25 brokers. Every asset class. One API.

Every tool callable via REST or MCP. Exact decimal precision on all financial values. Every call metered, audited, and encrypted. Built for machine to machine economy.

The layers

Six layers handle the full trading lifecycle.

Layer	What it handles	Count
Risk	Position sizing (G-formula), Iron Fist validation, kill switches, daily loss controls, portfolio risk	12 tools
Intelligence	Regime detection, pattern recognition, structural breaks, trend analysis, IV surface, options flow	11 tools
Analysis	Monte Carlo simulation, Kelly criterion, drawdown analysis, Sharpe/Sortino/Calmar, System R Score	18 tools
Planning	Options position sizing, futures position sizing, multi-leg strategy building, spread planning	4 tools
Execution	25 broker adapters across equities, options, futures, crypto, forex, prediction markets	25 brokers
Memory	Behavioral fingerprinting, bias detection, trajectory prediction, trade journal, anomaly detection	8 tools

Built on top of the layers

Every layer runs on shared infrastructure that enforces privacy, memory, auditability, and security by default.

Capability	Detail
Per-agent encryption	AES-128-CBC (Fernet) across all data paths. Strategies, credentials, and wallet data encrypted at rest. Only the owning agent can decrypt.
Compounding memory	Agents store observations, recall patterns, and detect their own cognitive biases. Behavioral fingerprinting and trajectory prediction improve with every session.
Complete audit trail	Every data read, every risk check, every decision logged. What was read, what passed, what was decided and why. Compliance ready.
Zero-trust architecture	Platform Guardian monitors all activity. Threat scoring, rate limiting, session binding, request signing. Honeypot routes detect probing.

Works with everything you already use

Connect via MCP protocol, Python SDK, or REST API. Bring your own market data from any provider.

Agent clients

Client	Integration
Claude Desktop	MCP config in `claude_desktop_config.json`
Claude Code	`claude mcp add systemr`
Cursor	MCP settings JSON
Any MCP client	SSE at `https://agents.systemr.ai/mcp/sse`
Python	`pip install systemr`
REST API	`POST /v1/tools/call` with `X-API-Key` header

Data sources

System R is data-source agnostic. You bring the market data, we do the analysis, risk management, and execution.

Provider	What it provides
Financial Datasets	Financial statements, SEC filings, earnings, insider trades, stock prices
Polygon	Real-time quotes, OHLCV bars, options chains, news
Alpha Vantage	Stock prices, forex, crypto, technical indicators
SEC EDGAR	10-K, 10-Q, 8-K filings (free public API)
Any OHLCV source	Any provider that returns open/high/low/close/volume bars

Quick start

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",
)
 
print(gate["approved"])     # True
print(gate["shares"])       # 363
print(gate["risk_score"])   # 23

One call. Position sizing, risk validation, and system health. Cost: $0.01.

Get started

Install the SDK, register an agent, and start calling tools.

pip install systemr

Register an agent and receive an API key
Call tools via REST (POST /v1/tools/call) or MCP protocol
Connect a broker and execute trades with built-in risk validation

Base URL

https://agents.systemr.ai

Authentication: X-API-Key header on all requests.