Quant macro analyst · Cloudfish × LuccAI

Ask the macro. Get the model's answer.

AlphaMacro is an AI analyst for markets that you talk to in plain English. It reads a quantitative model, refreshed every business day, covering equities, sectors, countries, rates, currencies, central banks and commodities, and every number it gives you traces back to the model.

Request early accessSee how an answer reads
1,252stocks with driver coverage
30countries
±100score scale
2016history from
Cross-asset scores · daily●
−100+100Illustration
You askedWhich assets flipped from bearish to bullish today?Three rows turned. Here they are, with the score change and what drove each one.

What an answer looks like

The call, the evidence, and the working.

AlphaMacro answers like a desk analyst writes a morning note. It leads with the reading, shows the numbers, cites where they came from, and suggests the next question.

> YOUHow is gold positioned?

Gold scores firmly positive, and the latest reading holds it there with a slight cooling.

Score now+46
Prior reading+41
Year change+23

The score has climbed for most of the year, driven by the macro components rather than price momentum alone. The latest reading is five points under the recent high, which keeps gold in the same band. A turn would need the macro inputs to weaken, and the model does not show that yet.

How I got thisRead the gold score history and the latest forecast vintage with query_score_history and query_forecast. Every result states its as-of date, vintage, source and row count.
Compare gold with silverWhat would flip this call?Show five years

Illustrative example. The figures show the format and are not a live reading.

  1. 01

    The call

    One sentence stating the model's reading, so you have the answer before the detail.

  2. 02

    The evidence

    The scores and chart behind the call, on the model's −100 to +100 scale, dated to the day they describe.

  3. 03

    The drivers

    Three sentences on what is moving the reading and what would have to change for it to turn.

  4. 04

    The working

    The tools that ran and the data they read, so any figure can be checked.

  5. 05

    The next question

    A comparison, a longer history, or the factors behind the move.

The product

A macro desk's toolkit, driven by questions.

AlphaMacro launches in stages. Each capability is marked with the stage it arrives in.

ASKAt launch

Ask

Plain-English questions about any market the model covers, answered as call, evidence and working.

CHATAt launch

Research threads

Longer conversations that keep context, with the source of every figure one tap away.

VIZAt launch

Charts

Score histories, comparisons, rankings and cross-asset heatmaps, exportable as images.

SCRNAt launch

Screener

Filter the 1,252 stocks with driver coverage by score, change, sector, country and market cap, then ask about the results.

RECPublic beta

Track record

Calls against outcomes, each call against what happened, with hit rates by conviction band and the sample size on every number.

ALRTPublic beta

Alerts

A watchlist that emails you when a score changes band or a reading turns, plus a weekly brief.

XRAYWith Pro

Portfolio X-ray

Upload holdings and see how the model reads each position and the book as a whole.

DECKWith Pro

Briefs and decks

Turn a thread into a branded PDF brief or slide deck for a client or an investment committee.

APIWith Pro

API and MCP

Query the same model from your own code or spreadsheet, or connect it to the AI assistant you already use.

How it works

A published quant model, and an analyst that only says what the model says.

STEP 1 · SCORE

The model scores

Each business day LuccAI's model scores each stock, sector, country and asset class from −100 to +100 and publishes that reading. Every vintage is kept, so what the model said at the time can always be shown.

STEP 2 · READ

The analyst reads

When you ask, AlphaMacro runs the analysis tools it needs against those scores. It does not invent figures. Each result states its date, vintage, source and row count, and the answer carries that through.

STEP 3 · ANSWER

You get the reading

A call, the evidence, the charts and what would change it. Take it further, save it to a brief, or set an alert on it.

The scale every answer uses

+46 gold (example)
−100−500+50+100
BearishNeutralBullish

What AlphaMacro will tell you

  • How the model reads a market now
  • How a reading has moved, and which inputs moved it
  • How past calls turned out, with the sample size stated
  • Percentage returns where the record supports them
  • When a question falls outside what the model covers

What it will not do

  • Tell you to buy, sell or hold anything
  • Give personal investment advice
  • Project dollar values, equity curves or a path the model did not publish
  • Optimise a portfolio for you
  • Present a guess as a model reading

Coverage

Top-down macro, bottom-up stocks.

Ask about a single company, a sector across regions, or how rates, currencies and commodities line up against each other today.

EquitiesSectorsCountry indicesGovernment bondsFXCentral banksCommodities
1,252stocks with driver coverage (Aug 2026)
30countries
11sectors
2016history begins, with every vintage kept

Track record

Published, not promised.

A forecast is only worth what its record shows. AlphaMacro publishes the model's calls against what happened, openly, for anyone to read.

  • CALLSEach published call against what happened, since the record starts
  • BY BANDHit rates for strong, moderate and weak calls, reported separately
  • SAMPLEThe count and start date printed beside every figure
  • REFUSEDNo rate shown where the sample is too small to mean anything
100%75%50%25%0%StrongModerateWeakn = …n = …n = …Layout preview. The live record publishes with the public beta; no figures are shown until then.

Pricing

Start free. Upgrade when the answers earn it.

Free$0

See how AlphaMacro reads the market, one month behind.

Join the waitlist
  • 3 questions a day
  • Top and bottom ten names
  • One regime reading a day
  • Readings delayed by one month
Explorer$49/ month

For individual investors who want the full model.

Request early access
  • 30 questions a day
  • The full universe, as of today
  • Score history and the latest readings
  • Charts and a watchlist with alerts
  • Weekly report by email
Pro · per seat$199/ month

For advisers and analysts who work with clients.

Request early access
  • Unlimited questions, fair use
  • Screener and Portfolio X-ray
  • Branded PDF briefs and decks
  • Track-record tools
  • API (10,000 calls a month) and MCP
Desk and EnterpriseLet's talk

For teams, wealth firms and platforms.

Contact us
  • Five seats and up
  • White-label briefs
  • History and vintage access
  • Licensing for platforms

Launch prices in US dollars, before tax. Every limit is shown in the app before you reach it, and nothing is charged on the Free plan.

About

Built by Cloudfish in collaboration with LuccAI.

The model

LuccAI

LuccAI builds and publishes the quantitative model: scores across stocks, sectors, countries and asset classes, refreshed every business day, and a record of every vintage it has published. The method is LuccAI's, and so are the numbers.

The analyst

Cloudfish

Cloudfish builds the analyst that reads the model: the AI agent, its tools, the app and the checks that keep every answer tied to a real figure. Cloudfish is an AI platform company based in Dubai.

Two teams, one rule: nothing reaches you that the model did not say.

Questions

Before you ask the model, ask us.

Is this investment advice?

No. AlphaMacro reports what a quantitative model reads about markets. It does not know your circumstances, it does not recommend buying or selling anything, and it is not a licensed adviser. Use it as research, alongside your own judgement or your adviser's.

Does the name mean it generates alpha?

No. Alpha is the language of the desks it serves, not a promise. The model publishes scores and calls; its record, hits and misses included, is public so you can judge it yourself.

How current are the readings?

Readings refresh every business day. Paid plans read today's reading; the Free plan reads one month behind. Every answer states the date and vintage it is based on.

Can the AI make numbers up?

AlphaMacro is built so that figures come from tool results, not from the language model's memory. Each result carries its source, date and row count, and the answer shows how it was produced. If the model does not cover something, AlphaMacro says so.

Can I use it from my own tools?

Pro seats include an API and an MCP connection, so you can query the model from a spreadsheet, your own code or another AI assistant.

Sign in

Already invited? Sign in here.

Accounts open with the private alpha. Until then, the app preview shows every screen on example data, labelled as such.

Open the app preview

We will email you a sign-in link. No password to remember.

Early access

The first readers are invited now.

A private group is using AlphaMacro daily before the public beta opens with the Free and Explorer plans. Tell us who you are and what you would ask first.

We use your email only to contact you about AlphaMacro.