Strip the mystique off a Bloomberg Terminal and it's four things stacked on top of each other. A data layer: prices, macro series, filings, holdings, news — collected from everywhere, cleaned, and delivered in one place. An analytics layer: the models and screens that turn data into judgment — what's trending, what's stressed, what's rich, what's abnormal. A news and narrative layer. And a network layer: chat, execution, the fact that every counterparty is also on it. The seat costs about $24,000 a year — roughly ₹20 lakh in India — whether you use all four layers or one.
Here's the thing that changed over the last decade, quietly: for daily-frequency work, the first three layers are now assembled from public sources. Not approximated — assembled. Exchange disclosure rules, central-bank data portals, regulatory filings and open-source models have moved most of the raw material into the open. What's left behind the paywall is real-time speed, exhaustive breadth, and the network. Those are worth $24,000 to an institution. They are worth very little to someone making daily or weekly decisions with their own capital.
The expensive part of a terminal is no longer the information. It's the speed, the breadth, and the chat. If you don't need those three, the rest can be free.
Regime Compass is the proof-of-work for that claim. Every tool on this site runs on data anyone can access, processed by models whose specifications — and whose audited failures — are public. The full stack, source by source:
| Terminal function | Free source | What we built on it |
|---|---|---|
| Prices, 20 markets | Yahoo Finance daily closes | 3-state HMM + SMA/EMA regime models, 0–100 risk gauge, regime-conditional movers |
| Systemic risk | same price panel, cross-asset | Turbulence (Kritzman & Li) · absorption ratio, market topology, correlation matrix |
| Macro | FRED, US Treasury, exchange data | Macro surprise meter, yield curves, growth×inflation quadrant |
| Valuation | Shiller's Yale dataset, NSE, CoinMetrics | CAPE, Nifty PE, MVRV — each as a percentile of its own history |
| Smart money | NSE bulk/block deals, EDGAR Form 4, STOCK Act filings, TWSE, IDX | Deal-level flow tracking across four markets, classified and netted daily |
| News sentiment | public RSS + FinBERT (open-source) | Every headline scored, daily tone gauge per market |
| Themes | constituent prices + Google News | 32 purity-weighted baskets with a cycle-stage classifier |
| Primary market (India) | aggregated grey-market quotes, exchange subscription data | Live GMP for every Indian IPO, calibrated against 268 actual listings |
Two things make this a terminal rather than a pile of dashboards. First, everything updates itself daily — the same discipline a Bloomberg imposes, without the seat. Second, the layers check each other: the regime verdict is only trustworthy because systemic stress, macro, smart money and sentiment are independent streams that can disagree with it. That cross-examination is the actual product; any one signal alone can lie.
An honest comparison has a second column. There is no free substitute for:
The line, then, is clean: if your workflow is execution, credit, or intraday, rent the terminal. If your workflow is judgment — what kind of market is this, how fragile is it, what is informed money doing — the free layer now covers it.
Partly because the marginal cost is near zero — the data is public and the models are code. Mostly because publishing forces honesty. A paid terminal never has to show you its accuracy; we publish walk-forward audits of our own models, including the finding that they don't beat buy-and-hold as return-timers — their value is drawdown behavior, not clairvoyance. A free tool that shows its losses is a better instrument than an expensive one that doesn't have to.
Open the terminal — today's verdict across all 20 markets →