Mcp

QuantRisk

Portfolio risk analytics — VaR, Monte Carlo, optimization, options Greeks, stress testing.

Verified: 2026-05-13 (mcp-registry-ingest-2026-05-13+enrich-capability-skill)

When to use QuantRisk

Choose if

You need real quantitative risk numbers (VaR, Monte Carlo, Greeks, optimization, stress tests) inside an agent loop — not a hallucinated estimate. Free tier (100 calls/day) is enough to validate the surface; $29/mo Pro tier unlocks production-grade simulation counts (100k Monte Carlo paths) and larger portfolios (500 positions). Stronger than asking the model to "estimate VaR" from a holdings table.

Avoid if

You need portfolio holdings ingestion, broker execution, or live market data — this MCP only computes analytics on positions you pass in. Also avoid if your agent calls risk analytics many times per minute on the free tier; the 100 calls/day cap will throttle quickly. For institutional use, the absence of a third-party math audit may be a blocker.

Risk Flags

  • MEDIUM cost Free tier is hard-capped at 100 calls/day, 20 positions per call, and 1 ticker × 1 year of price history — production agent workloads will exceed this quickly and require the $29/mo Pro tier (5,000 calls/day, 500 positions, 20 tickers × 5 years). Monte Carlo paths jump from 1,000 (free) to 100,000 (pro), which materially changes statistical validity.
  • LOW scope Risk analytics only — no portfolio holdings ingestion, no execution, no market-data subscription. Agents pass positions/prices in; the MCP returns analytics. Pair with a brokerage MCP or a portfolio holdings source.
  • LOW maturity Single-developer project (78degrees). Math layer described as "pure TypeScript, no external dependencies, fully unit-tested" — reasonable claim for a closed numerical surface, but no third-party audit disclosed.

Cost

Type: Freemium

Distribution

MCP Registry
dev.quantrisk/mcp-server
Repository
https://github.com/78degrees/mcp-server