// About

We build the mirror, you take the trade.

MirrorQuant is a local-first research cockpit for traders who want institutional-grade signal without giving up their data.

// Our story

From a frustrated trader's notebook to a research platform.

MirrorQuant started in 2023 when our founder, tired of stitching together Bloomberg exports, Polymarket screenshots and Python notebooks, decided to build the dashboard he wished existed.

The first version was a private tool — a BTC 5-minute prediction engine paired with a stock factor radar. After a year of running it on live capital, we opened it up to other quants who wanted the same edge without rebuilding the plumbing.

Today MirrorQuant unifies prediction markets, equity intelligence, macro context and AI research in a single, local-first cockpit. Your strategies, trades and market data never leave your machine — we only ever verify your license.

Share this story
Share

Mission

Make institutional-grade signal accessible — and private.

Give independent traders the same quality of cross-asset signal that a discretionary desk has, while guaranteeing local-first privacy.

Vision

One cockpit for every directional bet you make.

Whether you trade event probabilities, single names, or macro themes — the same data layer, the same AI, the same review loop.

// Why this platform

Why prediction markets + stock radar?

01

Prediction markets price reality first.

Event markets often lead headlines by hours and bake in conditional probabilities a pure technical view can't see.

02

Stocks turn that view into P&L.

A multi-factor radar (value, growth, momentum, smart-money, sentiment, quality) maps a macro thesis to actionable single-name baskets.

03

AI compresses the read.

Daily AI briefs translate noisy data into a 30-second thesis, with an explicit confidence and dissent view.

04

Local-first is non-negotiable.

Strategies and trades stay on your machine. We only verify your license key — never your positions.

// Methodology

How we build signal.

Hypothesis-first

Every model starts with a falsifiable thesis (e.g. 'rate-cut repricing rotates into industrials'), not a black-box fit.

Walk-forward backtests

All factor scores and AI calls are evaluated with strict walk-forward and out-of-sample windows; no in-sample peeking.

Multi-source fusion

We blend market data, order-flow, on-chain, fundamentals, prediction-market odds and sentiment — each weighted by its measured edge.

Calibrated, not loud

We report probabilities and confidence — not 'BUY NOW' — and publish hit-rates so you can size against measured edge.

// Data sources

Where the numbers come from.

  • US equities: consolidated tape (SIP) for prices, vendor fundamentals.
  • Prediction markets: Polymarket, Kalshi public APIs.
  • Macro: FRED, US Treasury, ECB, BLS, BIS.
  • Crypto: major spot exchanges + on-chain (Glassnode-grade aggregations).
  • Sentiment: news + social, scored by our in-house LLM pipeline.
  • Smart money: 13F filings, ETF flows, dark-pool prints.

// Team

Built by traders, for traders.

Alex Chen

Founder · Quant Research

Ex-hedge-fund quant; 8 yrs in derivatives & factor modeling.

Mira Park

Head of AI Engineering

Focused on LLM-driven structured extraction & backtests.

Jordan Liu

Full-stack & Infra

Builds the local-first execution & backtest engine.

// Risk disclosure

Read this before you trade.

MirrorQuant is a research and analytics tool. Nothing on this site is investment advice. Markets carry risk of substantial loss. Backtested or simulated results do not guarantee future returns. Always size positions to risk you can afford to lose.

Full terms & disclaimers