Concordal
The Decision Dialectic · Concord
牌照申请筹备中 · HK SFC Type 4 · 多 agent 辩证 · 完整 audit log

The dialectical, auditable AI advisor.像辩论员而非黑盒的 AI 投顾.

The bull

“Buy.”

多头:买入。

Services revenue is compounding at 14% year-over-year and now comprises 26% of the top line. Margin mix shift is structurally under-priced.

The bear

“Sell.”

空头:卖出。

At 28× forward earnings the structural story is fully priced. Vision Pro is soft. China revenue is decelerating. Wait for the next quarter.

The manager · synthesises

Reduce both sidesto one trade.

把两边的话,化成一笔交易。

Seven specialist agents — fundamentals, sentiment, news, technical, macro, plus bull-and-bear advocates — argue in writing. A manager weighs every line and outputs a single, confidence-calibrated call. The whole transcript is auditable.

No password · email later · 90 seconds

27
Regression tests · zero lookahead
2
LLM providers live · DeepSeek V4 + Gemini
3
Markets · US · A-share · Crypto
What brings you here?

Four jobs. Decide, browse, verify, build.

Each card mirrors a navigation group above — every page is within 3 clicks.

Decide

Run a 7-agent debate on any ticker

Bulls and bears write opposing reports. Risk panel votes. Manager signs off. Every LLM call traced.

Markets

What's interesting today

Limit-up pool, fund flow, sectors, HK, ETF, calendar — A-share + worldwide in one place.

Proof

Does it work? Backtest evidence.

20 ticker × 78 week backtest — sharpe + hit rate. Click any decision to trace LLM reasoning.

Research

Power-user research tooling

FRED → Qlib → Backtrader → Lean live chain. Ecosystem map shows 12 integrations in real time.

跨市场统一覆盖 · 论文 §10.2

AAPL、600519、BTC/USDT — 同一个 7-agent 流水线、同一个 UI、同一个置信度解读。散户今天有三个不相交研究栈(彭博美股、东方财富 A 股、Coingecko 加密币), Concordal 统一覆盖让你应用一致决策准则、消除跨市场认知负荷。

Why role separation

One ChatGPT prompt cannot do five specialists' jobs.

一条 prompt 干不了五个专家的活。

Ask a single model “should I buy AAPL?” and it hedges every direction. Attention is finite. Conflicting signals get smoothed into a HOLD with low confidence — rarely the optimal trade.

Our pipeline runs five specialist analysts with separate prompts, separate context windows, separate evidence. Each forms an opinion in isolation. Then a bull and a bear persona read all five reports and write opposing pitches. Then a trader synthesises. Then risk approves. Then a manager signs off.

In our 78-week backtest across 20 tickers, the multi-agent pipeline produced calibrated confidence — the system's 70%-confidence calls were right roughly 70% of the time. The single-prompt baseline was systematically over-confident.

Architecture

Seven stages. Every one auditable.

七个阶段。每一步都可回溯。

  1. 01

    Fundamentals

    SEC EDGAR · akshare

  2. 02

    Technical

    OHLCV · Alpha158-lite factors

  3. 03

    Sentiment

    Reddit · 东方财富股吧 · 雪球

  4. 04

    News

    Reuters · WSJ · Bloomberg wires

  5. 05

    Macro

    FRED · OpenBB

  6. 06

    Bull / Bear debate

    Two adversarial personae

  7. 07

    Manager + risk

    Synthesis + position size + stop

Coverage

Three markets. Five analyst lenses. Real data, every cell.

三个市场,五个视角,每一格都是真实数据。

MarketFundamentalsTechnicalSentimentNewsMacro
US Equity
yfinance · SEC EDGAR XBRL
A-Share
akshare · 东方财富 · 雪球
Crypto
CCXT · Binance default

A Bloomberg seat costs $25,000.
The reasoning behind a good trade should cost cents.

一台 Bloomberg 终端两万五;
一次好的决策推理,应该只值几分钱。

— Concordal · 2026

See your first decision in 90 seconds.

90 秒,看到你的第一份决策。

Magic-link sign-in. No password, no card. First decision is free. Authenticated users get five real-LLM decisions a day.

Have an invite code? Redeem →

Decision support · not investment advice · markets remain uncertain · full disclaimer