Shipped: hpsilab-mcp — Python SDK for HPSILab quant tools
Open-source (MIT) Python client for the HPSILab API: pip install -U hpsilab-mcp
One client gives your scripts, backends and agents the same 9 tools as the HPSILab MCP server:
- analyze_stock: one JSON analysis with direction, score, bullish and bearish factors and a plain-English summary
- get_pretrade_risk_scan: volatility, beta, drawdown and VaR vs SPY, with fixed pass/warn/fail limits
- get_iv_radar and get_option_pressure: IV regime, skew, gamma wall, call/put walls, max pain
- get_monte_carlo: 10-day price range and odds of finishing higher
- plus AI prediction, backtested equity curves, chart images and research reports
Python 3.9+, typed errors (HpsiMcpError). Register at hpsilab.com/ and verify your email for 100 trial Credits. Agents that prefer MCP can connect to hpsilab.com/mcp instead.
Comments2
Giving 100 trial Credits after email verification is a simple way to get people trying it. Of the people who sign up, how many actually run a first analysis, and which of the nine tools do they reach for first? I'm curious whether the pre-trade risk scan or the stock analysis pulls people in.
Thanks Ryan. Still early, so I don't have an activation rate worth quoting yet. The design bet is the pre-trade risk scan: it's the one answer an agent can act on directly (pass, warn or fail against fixed limits), while analyze_stock is the easier first call because one request returns everything. Which of those two people come back to is exactly what I want to learn next. hpsilab.com/ — Haiyun
2 comments
Giving 100 trial Credits after email verification is a simple way to get people trying it. Of the people who sign up, how many actually run a first analysis, and which of the nine tools do they reach for first? I'm curious whether the pre-trade risk scan or the stock analysis pulls people in.
Thanks Ryan. Still early, so I don't have an activation rate worth quoting yet. The design bet is the pre-trade risk scan: it's the one answer an agent can act on directly (pass, warn or fail against fixed limits), while analyze_stock is the easier first call because one request returns everything. Which of those two people come back to is exactly what I want to learn next. hpsilab.com/ — Haiyun