Explore economic conditions, market trends and model research in one place. Compare the evidence, understand its limits and investigate your own investment questions.
A pattern that described the past may fail later. We examine model outputs alongside dated outcomes, economic context and competing explanations. Retraining alone does not establish an investment advantage.
Check whether a result persists across different market conditions. A model fit or a high stored score is not a probability of a profitable holding-period return.
Examine losses, missing observations and alternative explanations alongside promising results. Choosing only successful examples can make weak research look convincing.
For tariffs, inflation and interest rates, distinguish announcements from effective dates and ask what was already expected. An important event does not establish a predictable market response.
Choose the format that fits your investigation.
Read sourced news, reviewed economic events and aggregate model results. Our research brief helps you form questions and weigh the evidence without a daily list of trades to copy.
Use the REST API to investigate market data, source context and stored model diagnostics. Check timestamps, coverage and methodology before comparing results. Start with a free key, with no card required.
Connect a research assistant through MCP to retrieve data and compare sources. Keep decisions under your control; model-generated interpretations need verification and are not instructions to place trades.
Use MoneySignals to organize your research. The decision remains yours.
Check dates, coverage and provenance. Distinguish an observed fact from a source's interpretation or a model estimate.
Read the complete record, including losses. Compare the same dates, costs and exposure; daily direction accuracy is different from buy-and-hold returns.
Ask what would invalidate it and whether enough independent evidence exists. An inconclusive result can be useful information.
Research in development: we are building 5–10 and 20–60 trading-session comparisons. Validated longer-horizon findings and profitable dip timing are not available product claims. An edition may contain no qualified finding.
Explore event-driven market research and inspect how earlier signals resolved. Detailed outcomes, methodology and the signal ledger are available below. These historical direction results do not establish longer-horizon portfolio returns.
Loading recorded outcomes; no result is assumed while data are unavailable.
Sum of direction-signed per-signal returns — a correct bearish call counts positive (equal-weight, not compounded)
All reported types, including weak results. Counts are resolved signals, not independent trials. A 50% coin flip is only one comparison; class balance, shared market moves, costs and benchmark returns also matter.
Live data from GET /track-record — an outcome is recorded only after a signal's time horizon has passed, so listed signals are no longer actionable. Signals without a recorded outcome are not scored as wins or losses. The displayed sum uses direction-signed per-signal returns (bearish returns negated); it is not a funded portfolio, and it does not account for trading costs or overlapping capital. Past performance does not guarantee future results. Not investment advice.
Inspect an ETF pullback study, from its fixed rules to the recorded outcomes and limitations. Historical backtests are separate from live results.
This archived study examined fixed holding windows after ETF pullbacks within an uptrend. The figures shown are a selected four-fund subset that passed the earlier study gate, not the complete candidate population. They do not establish the performance of our service or qualify a new longer-horizon finding. Review unsuccessful periods and the full research record before interpreting this example.
Backtest statistics, not live results. The four ETFs are correlated — simultaneous fires are closer to one bet than four. The 2005–2014 era (which contains 2008) did not clear the gate. The fixed short hold is unprotected by design: every canonical condition exit, stop and profit-target overlay was tested paired on the same fires, and all of them cost expectancy. Fires are rare (~8–10 per year per ETF). Not financial advice.
Historical research diagnostics and methodology for all 13 ETFs, including screening failures. No active windows or entry instructions:
GET /api/v1/signals/etf-pullback
or the get_etf_pullback_signal MCP tool — included with the free key.
Read the docs →
The figures above are an archived July 2026 snapshot of a selected four-fund subset, using dividend/split-adjusted history since 1993. The API reports dated historical measurements for the full 13-ETF study universe, with losses and unavailable histories retained. A new computation does not establish fresh source coverage or a qualified investment edge.
One plan — every tier includes both the daily email and API access
News coverage depends on source availability and review. Unavailable coverage is identified in the brief; paid access does not guarantee a number of headlines, qualified findings or better investment returns.
ECB source releases are also available free from the European Central Bank.
Already hold an API Starter or API Pro key? It keeps working unchanged — those plans map to Premium and Professional.
Discuss API volume, data feeds and support for your platform. The options below require a separate agreement; they are not included by default in a subscription.
The world briefing and news map are free to read, with no account. Sign in to add your portfolios and economic data. Source dates and coverage appear alongside the data.
Connecting an account supports portfolio visibility; it does not establish the quality of a model or authorize automated investment decisions.
Straight answers before you sign up
No. MoneySignals is for informational and educational purposes only — data, statistical models, and automated analysis, not investment, financial, legal, or tax advice, and not a recommendation to buy or sell any security. You are always responsible for your own decisions. See our Terms of Service.
Read the recorded direction results, including weak periods and losses. A daily hit rate does not tell you whether a weekly or monthly holding strategy earns a positive return after costs. We are developing those comparisons; we do not currently claim validated longer-horizon returns or reliable dip timing. Model scores are not established probabilities of profitable returns.
The legacy event-signal record reports direction outcomes and price moves under its stated measurement rules. It is separate from actual broker fills and from the new buy-and-hold research accounting. Summing direction-signed signal returns is not a cumulative account: costs, overlapping positions and capital requirements matter. Inspect dates, pending outcomes and methodology alongside the numbers.
Free includes a research brief with a business-news preview, reviewed events and aggregate results, plus a free API key. Premium adds broader news and API research coverage. Professional adds deeper news/calendar coverage, batch diagnostics, social data and higher API limits. Paid plans provide more data and tools, not higher promised win rates or earlier trades. See Pricing.
Join the free daily email and click the confirmation link — your key arrives in the welcome email (5 requests/min, 100/day). It works on both the REST API and the MCP server for predictions and market context. Lost it? Sign up again with the same email and confirm: the old key is revoked and a fresh one is emailed to you.
Yes. Paid plans are month-to-month with no contract — cancel anytime and you keep access through the end of the billing period. Manage or cancel from the Stripe billing portal linked in any receipt.
End-of-day market data from EODHD, news and economic-calendar events, and alternative data we collect ourselves — social trends (Reddit, YouTube), Senate lobbying disclosures, and congressional trades, plus experimental "possible signal" feeds like DC-area corporate flight tracking that we surface as-is rather than fold into the models. It's combined into the daily reads.
The research brief is scheduled for trading days. Source and model refresh times vary; check their timestamps and coverage rather than assuming everything was refreshed overnight. Daily data collection does not mean a new qualified research finding every day.
Payments are processed by Stripe — we never see or store your card number. We collect account and service-usage data, plus the portfolio information you choose to connect or submit. We don't sell personal data. See our Privacy Policy.
Yes, within your plan's rate limits. Bulk redistribution or reselling of the raw data and signals needs prior written permission. See the Terms of Service for the details.
Use REST and MCP to retrieve market data, source context and clearly labeled model diagnostics for your own analysis. Research briefs summarize evidence and limitations; ranked trading signals and action-generating trade plans have been retired. Explore the full documentation →
Connect Claude — Desktop, Code, or any MCP client — to inspect market context, source coverage and recorded model outcomes. The track record needs no key; a free-signup key unlocks model diagnostics, market context and the historical ETF study. Model scores are not calibrated probabilities of profit, and no qualified holding-period strategy is established. See the research workflow.
claude mcp add --transport http moneysignals https://api.moneysignals.us/mcp \
--header "X-API-Key: at_sk_your_key"
Prefer a local process? A one-file stdio server is a curl -O away — both options in the docs.
Click any button to hit the real API — no signup required
{ "status": "ready" }