Skip to content
Live

Lucrum — AI Quantitative Trading Platform

Describe strategies in natural language — AI generates vnpy code and backtests · 11 investment advisors

11Investment AdvisorsMulti-perspective Agents
30+Backtest MetricsSharpe / Drawdown / Win Rate…
3157+Test CasesVerified by Vitest
Decimal.jsPrecisionZero floating-point error

What is Lucrum?

Lucrum is the AI-Native quantitative trading decision platform from Lurus. The core idea: natural language is the best programming language — describe your strategy in plain English, and the AI automatically generates code, runs backtests, and evaluates across multiple dimensions. It ships with 11 professional investment-advisor Agents (perspectives such as Buffett / Peter Lynch / Livermore / Simons) and platform-wide financial-grade precision computation via Decimal.js (verified by 3,157 Vitest test cases) with zero floating-point error.

The name comes from the Latin "Lucrum" (profit), evoking precise insight into market opportunities.

Current stage: public beta

lucrum.lurus.cn is live, and the core features (strategy generation / backtesting / AI advisors) are ready to try directly; see pricing at /pricing. It has not yet reached general availability (GA), and some advanced capabilities (strategy marketplace, live brokerage integration) are still being refined.

Core Capabilities

From a single English sentence to a rated backtest

Strategy generation, multi-agent research, strategy marketplace, quota billing, real-time execution — one end-to-end pipeline.

AI Strategy Generation & Backtesting

Natural language → code

Describe your strategy intent in plain language, and the AI automatically generates vnpy CtaTemplate strategy code.

Financial-grade backtesting

Full Decimal.js precision, A-share lot constraints (multiples of 100 shares), T+1 rules, commissions + stamp duty + transfer fees + slippage.

30+ metric analysis

Sharpe ratio, maximum drawdown, Sortino, Calmar, win rate, profit/loss ratio, and more.

S/A/B/C/D five-tier rating

Weighted across 4 dimensions: returns 30% + risk control 30% + stability 25% + efficiency 15%.

11 AI Investment Advisors

A multi-agent investment-analysis system orchestrated with LangGraph (4 analysts + 2 researchers + 4 masters + 1 debate moderator = 11):

4 investment-master perspectives

Buffett (value), Peter Lynch (growth), Livermore (technical), Simons (quant).

4 analysts

Fundamentals / technicals / sentiment / macro, each delivering its own conclusion.

2 researchers + 1 debate moderator

A Bull vs. Bear debate that avoids single-perspective bias.

Memory-engine integration

Remembers your trading preferences and past decisions via MemX.

Strategy Marketplace

An open quantitative-strategy ecosystem that connects strategy developers and traders:

RoleFeatures
Strategy authorUpload strategies, set prices, and view revenue splits
Strategy userBrowse and subscribe to strategies, deploy to live trading in one click

Revenue split: platform 30% / strategy author 70%.

Quota & Billing

  1. Plan limit — the monthly AI call allowance included in your subscription plan.

  2. Redis monthly counter — tracks this month’s usage in real time.

  3. Lubei balance fallback — once the quota is exhausted, charges are automatically deducted from your Lubei wallet; 1 Lubei = 10,000 tokens.

Real-Time Data & Execution

Market coverage

A-shares (Shanghai & Shenzhen exchanges, ~5,000+ stocks, data sources adata + Eastmoney); Hong Kong / US / crypto planned.

Simulated trading

A built-in Mock Broker that fully simulates T+1 rules, 100-share lots, commissions, and stamp duty.

Risk-control engine

Position limits, stop-loss / take-profit, and maximum-drawdown protection.


Who It’s For

Run quant without writing code

User typeHow Lucrum helps you
Quant beginnersThe AI assistant guides you through onboarding — just describe your strategy in plain language to generate a code skeleton
Individual investorsPick validated strategies from the marketplace and deploy them in one click, no coding required
Strategy developersA complete develop-backtest-launch toolchain; list your strategies and earn passive income
Professional trading teamsAPI endpoints to integrate with your existing trading systems

Comparison

How it differs from traditional quant platforms

Compared to traditional quant platforms

维度Lucrumvnpy掘金米筐聚宽
Strategy authoringNatural-language generationHand-written PythonHand-written PythonHand-written PythonHand-written Python
AI investment advisors11 multi-perspectiveNoneNoneNoneNone
PrecisionFull Decimal.js precisionfloatfloatfloatfloat
Strategy marketplaceBuilt-in + ratedNoneYesYesYes

Technical Architecture

From browser to settlement engine

Lucrum layered architecture

Next steps

Powered by Lurus Technologies