Where Macro Meets the Price Chart
Trade with context, not just candles. CPI prints, Fed decisions, news tone and geopolitical escalation on the same axis as gold, BTC or the S&P — on your own machine, with your own keys.
Hand the research to it. A macro analyst sits inside the terminal: it watches the relationships between your series, runs the event studies, and tells you when one breaks — and, most days, that nothing happened. See it working →
Not on sale yet · planned pricing · nothing is chargedMacro on the Price Axis
- Overlays: CPI, Fed Funds, 2Y/10Y, M2, unemployment and PMI normalised onto your chart — shifted to their release date, so an August print appears where the market saw it
- Events: FOMC decisions, CPI prints, ACLED escalation spikes and GDELT tone shocks as tagged vertical markers, each carrying the 5 / 20 / 60-session reaction history for that asset
- Correlation: a rolling 30 / 90 / 250-session panel between the asset and every active overlay, with lead / lag detection — on gold, CPI YoY reads r = 0.65 at 90 sessions and 0.51 at 250, while ACLED escalation runs the other way round at 0.29 and 0.49. A number, not a hunch — and the two windows disagreeing is itself the finding
Events on the Axis, With What Followed Before
- Real-time event feed: 15-minute GDELT slices, daily ACLED aggregates, FRED releases the minute they publish and Friday's COT — one stream, every source tagged
- Base rates, not signals: ACLED escalation crossing 70 on MENA — Brent's median path over the next 20 sessions was +3.2 % across the 15 comparable crossings in the ten-year cache, and 3 of those 15 went the other way, the worst of them −7.7 %
- Transparency: every marker links to the event study behind it; every index in the feed is recomputable from the public source with your own keys
An Educational Approach
- Expertise: the Weekly Brief turns the engine's own numbers into event studies, how-tos and opinion — written by the RegimeLens Desk, not generated from headlines
- Measure, don't feel: a correlation you can read beats a macro opinion you cannot falsify; the panel shows 30, 90 and 250 sessions side by side so you see when a relationship is recent, persistent or fading
- Method: release-date alignment, log-return correlation, cooldown-filtered event studies — every recipe is three lines of YAML you can read, edit and export
Hand the Research to It
You state what you want watched, once, in a sentence. After that the terminal works on its own clock: it checks, it acts, it writes down what it did and the number that caused it — and on the checks where nothing crossed anything, it says so instead of inventing news.
Watch the relationships, arm what this screen implies, and brief me on a cadence
Chosen in words from four options in the terminal's own setup, on XAUUSD. Pause, undo the last action, change the instruction and stand the desk down are one click each in the panel header.
What it did, unasked
The first three entries in the standing desk after that sentence. Every line is written by the desk itself, with the reading that produced it.
- STANDINGStanding instruction set: Watch the relationships, arm what this screen implies, and brief me on a cadence It runs every 6 seconds on this terminal's clock, on XAUUSD. Pause, undo and change are one click each above. Three jobs at once means three sources of interruption. It will speak more often than any single instruction would, and some of it will be about things you did not ask about.
- ARMEDArmed CPI YoY above 3.2 % half a point above where CPI is printing now. It reads CPI YoY 2.7 % now, so it is false, and it crossed 11 times in the ten-year sample, last on 05 Nov 2024. You can see it and disarm it in the alerts panel.
- NOTHINGNothing worth telling you about Checked 6 things: MENA escalation 88 (moved 0 of 15) · CPI YoY 2.7 % (moved 0.0 of 0.15) · INFLATION tone 0.9 (moved 0.0 of 2.5) · strongest link still CPI at +0.65 (drifted 0.00 of 0.10) · 1 armed condition unchanged · liquidity regime still Easing. Nothing crossed a threshold I hold, so there is nothing to say. On a macro desk that is the honest answer most days.
And one thing it produced
The event study it opens when you ask what a CPI surprise usually does to gold. A CPI surprise is a print 0.2 points or more away from the one before it — a rule over the series, not a hand-picked list.
| Where it did not work · 26 of 61 | ||
|---|---|---|
| 18 Sep 2021 | CPI 5.7 → 6.2 % | −8.4 % |
| 18 Jun 2021 | CPI 5.1 → 5.4 % | −7.0 % |
| 18 May 2023 | CPI 4.4 → 3.9 % | −4.6 % |
| 18 Sep 2017 | CPI 2.0 → 1.6 % | −4.5 % |
| 18 Jan 2026 | CPI 2.4 → 2.7 % | −4.4 % |
| … and 21 more, every one of them drawn in red on the fan | ||
Our Products
Research firms stop at the report. RegimeLens ships the report, the tooling behind it and the portfolio framework that uses it.
The RegimeLens Terminal
A charting terminal with a macro brain. Runs on your machine, pulls FRED, GDELT, ACLED and CFTC through your own keys, bought once.
24 overlay recipes
CPI, Fed Funds, yields, M2, real yields, tone, escalation, positioning — plus your own in three lines of YAML.
Event markers & studies
Scheduled and derived events on the chart, each with the median / quartile path across every comparable instance.
Correlation panel
Rolling r on log returns vs overlay shocks, three windows, lead / lag from −10 to +10 sessions.
Local, BYOK, exportable
Keys encrypted on your disk, cache works offline, every series exports as PNG, CSV or Parquet.
The Weekly Brief
One event study, one how-to or one opinion piece at a time, written from the engine's own numbers. The archive is open.
Event studies
What Brent, gold and wheat did after 31 escalations; what SPX did around 214 breakouts with a CPI print inside the window.
How-tos
Getting a FRED key, writing a recipe, aligning to release dates, normalising honestly — with the exact files.
Chart of the week
One chart, one paragraph, the recipe to reproduce it in your own terminal.
No signals
Base rates and context, never buy / sell calls, never leverage. Educational by design.
Model Portfolios
Three diversified allocations, one per macro regime, with the weights set by the same overlays you see on the chart. Fully transparent, rebalanced monthly, sized conservatively.
Regime-driven
Real yields, CPI trend, Fed Funds direction and the escalation index decide which of three allocations is current.
Diversified by design
Equities, duration, gold, commodities, cash and a small digital-asset sleeve — no single position above 25 %, no leverage.
Drift tables
Target vs current weight, drift since last rebalance, and the rule that changes a weight when the regime flips.
Educational
A framework to study and adapt, not a recommendation. Export the weights and test them in your own tools.
Memberships
One licence you own, one feed you can cancel. Nothing else. Early access opens in stages.
Terminal
For traders who want the tooling on their own machine, once — and their keys, cache and charts to stay there.
Perpetual licence · 3 devices · after year one, keep the version you have or add updates for €99 / yr
Benefits:
- 24 overlay recipes + custom YAML recipes
- FRED, GDELT, ACLED and COT through your own keys
- Event markers, central-bank calendar, event studies
- Correlation panel with lead / lag detection
- Model-portfolio weights and drift tables in the terminal
- Export overlays as PNG, CSV or Parquet
- Weekly Brief archive and chart book
Runs on Windows, macOS, Linux or Docker · guest demo in the browser
Terminal + Macro & Sentiment Feed
For traders who want our indices, alerts and the live regime read refreshed on our servers instead of their laptop.
Everything in Terminal, plus the feed · cancel any time · the terminal stays yours
Benefits:
- GDELT tone index by theme, refreshed every 15 minutes
- ACLED escalation index for 8 regions, daily
- COT positioning z-scores for 15 markets, weekly
- Threshold alerts to Telegram or e-mail
- Live regime read and monthly model-portfolio rebalance notes
- Update contract for the terminal included while active
Access via the desktop app and the browser terminal
Planned terms: the feed renews monthly and can be cancelled at any time; the terminal licence is perpetual.
Research Coverage
US
Markets
- S&P 500, Nasdaq 100
- US 10Y, 2s10s, real yields
- CPI, PCE, payrolls, ISM
- Fed Funds & FOMC calendar
Global
Markets
- DAX 40, Nikkei 225, EM
- ECB, BoE, BoJ decisions
- Eurostat HICP, PMIs
- GDELT tone in 100+ languages
Commodities
Market
- Brent, WTI, natural gas
- Gold, silver, copper
- Wheat, corn
- ACLED escalation by region
Fixed Income
& FX
- Treasuries, Bunds, JGBs
- Credit spreads, TIPS
- EURUSD, USDJPY, DXY
- CFTC COT positioning, BTC
About
RegimeLens
RegimeLens is a macro-research terminal from Mast. We build the terminal we wanted to research with: macro data, geopolitical events and news sentiment aligned to the price chart, on your own machine.
We publish the Weekly Brief, maintain three regime-driven model portfolios and ship the tooling that produces both — because a chart of what happened is not enough; you need to know why, and whether it has happened before.
See It Working
Research firms describe the chart. This is the chart. Scroll to zoom, drag to pan, hover for the readout, toggle an overlay — then open the full terminal.
Overlays that line up
Monthly series are shifted to their release date and normalised over the window you are looking at — never a second axis you can fiddle with.
Events with a base rate
Hover a marker in the full terminal and you get the median and quartile path for that asset across every comparable event in the cache.
Yours, locally
The installed engine calls FRED, GDELT, ACLED and your exchange from your machine. Keys are encrypted on your disk; the cache works offline.
One Question. Then It Sets Itself Up.
The terminal opens by asking "What are you here to do?" One answer sets the instrument, the overlays, the correlation window and what it watches for you. It then says what it chose, why, and what the choice costs — because none of these choices is free.
-
"I trade gold and want the macro behind it"
What it chose: XAUUSD over three years, the correlation matrix on a 250-session window, FRED CPI YoY on the price axis because it is the strongest of the five at r = 0.51, geo-risk as the regime axis because gold is the instrument where that split is largest, and a standing instruction watching all of it.
What it costs: a 250-session window is a structural read. It will not notice a regime change for months. The 30-session one notices in weeks and is mostly noise — it picked the slow one on purpose, and it is the wrong choice if what you want is to react.
-
"I want to know what this week's data means"
What it chose: the forward calendar for XAUUSD and the event study cut on Escalation crossing, because that is the class of the next thing on the diary — ACLED file refresh in 3d. It has 20 past instances behind it. Plus a standing instruction that rebuilds the briefing every 8 checks, with what changed at the top.
What it costs: a calendar tells you what is scheduled, not what is priced. In this ten-year sample the largest single-session moves cluster around escalation crossings, which are never on anybody's diary — watching the week means not watching those.
-
"I want to be told when a relationship breaks"
What it chose: the geo-risk regime split, because across the three axes it is the one where the correlation moves most — +0.24 to +0.56 against +0.42 for the whole ten years, a 0.32 spread. It armed r to escalation below 0.00, which crossed 0 times in the sample, and set a standing instruction to tell you when the ranking or the number moves.
What it costs: a sign change is a narrow watch. ACLED can decay from +0.56 to 0.05 — a relationship that has completely stopped working — without ever crossing zero, and this condition would never fire. Widen it and it fires on noise instead. There is no setting that catches both.
-
"I want a briefing before the open, without asking"
What it chose: the macro overlay layout so the chart stays readable behind it, and a standing instruction that rebuilds the briefing every 8 checks. Every rebuild opens with what changed since the last one, computed against the exact readings the previous one was built from — so the first thing you read is the diff, not the whole page again.
What it costs: the brief is the same length on the mornings nothing happened. It will keep telling you nothing happened. That is the honest answer most days and it is also the price — you will read a page to learn it, and the day you stop reading it is the day it matters.
-
"I am just looking around"
What it chose: nothing. The chart stays exactly as it booted — XAUUSD over 252 sessions, one overlay on the axis, markers off — and it tells you the strongest of the five macro links is FRED CPI YoY at r = 0.65.
What it costs: nothing is set, so nothing is watching. You will only ever find what you go looking for, and the two things this terminal is actually for — the event study's failures and a condition armed while you are elsewhere — are both things you have to ask for.
Every answer is reversible: LAYOUT RESET puts the workspace back, the standing desk stands down in one click, and anything it arms is disarmed with a switch. It is a research terminal — it will not size a position or name a return.
Where It Did Not Work
Ask the analyst is this real? and it computes the spread of every instance and the interval around the median, from the same array the fan is drawn from. On XAUUSD, four of the five event classes it ships with produce a median that sits inside its own error bar. It leads with that.
| Event class | n | Median +20d | Went that way | Interval around the median | What the terminal says |
|---|---|---|---|---|---|
| CPI surprise | 61 | +0.3 % | 57 % | −1.0 % … +1.6 % | contains zero 26 of 61 went the other way |
| FOMC move | 29 | +1.9 % | 72 % | +0.2 % … +3.6 % | excludes zero 8 of 29 went the other way |
| Escalation crossing | 20 | −0.9 % | 50 % | −3.4 % … +1.6 % | contains zero 10 of 20 went the other way |
| Positioning extreme | 15 | +0.1 % | 53 % | −2.9 % … +3.1 % | contains zero 7 of 15 went the other way |
| Quarterly expiry | 40 | −0.1 % | 53 % | −1.9 % … +1.7 % | contains zero 19 of 40 went the other way |
XAUUSD, ten-year sample. Interval = median ± 1.96 standard errors. A median inside its own error bar means the class does not tell you a direction on this sample — use the study to see the spread, not to pick a side.
That is a description of 61 past instances on sample data. It is not a forecast, and 61 observations is far too few to trade a distribution from.Event study panel, under the fan
Nothing crossed a threshold I hold, so there is nothing to say. On a macro desk that is the honest answer most days.Standing desk, on a check where nothing moved
No condition here has ever been tested for whether acting on it made money, and I am not going to imply it has.The analyst, grading your armed conditions
Where Macro Meets the Price Chart
Open the guest terminal now, and join early access to plot FRED, GDELT and ACLED on your own charts.
Planned pricing: buy once, cancel the feed at any time. Not on sale yet.
Let's get in touch.
Write to us about early access, data sources or anything else on this site.
Or write to us directly at hello@regimelens.com. The button above opens your own mail app; nothing is sent from this page.