Lexfi is now available on the OpenBB App Marketplace. In this post, we walk through how Earnings Intelligence, Central‑Bank Insights, and Sentiment & Signals turn calls, press conferences, and news into daily indicators you can actually trade on, then show two real workflows running inside the OpenBB Workspace.
An equity analyst covering a sector reads dozens of earnings calls a quarter. A macro PM watches fifteen central banks on fifteen different calendars. A quant book wants both, joined to prices, going back a decade.
The information is there, but it arrives as prose, at speed, in language built to be parsed by hand.
Reading it all is slow. Comparing it is impossible. Off-the-shelf sentiment scores are too coarse to trade on, headline-driven news feeds miss the nuance that lives in the Q&A, and desks that build their own NLP end up with brittle pipelines that only cover one geography.
The job the user is trying to do, position around what management or a central bank actually said, before the market has fully digested it, never quite gets done at the breadth or the speed the desk needs.
Lexfi solves the extraction problem.
Lexfi processes earnings calls, central-bank press conferences, macro news, and social with proprietary language models, and hands back structured numeric indicators. The same schema for every company, every central bank, every quarter, every market. Once tone is a number on a date, you can join it to prices, backtest it, and trade against it.
Inside OpenBB, Lexfi’s indicators turn that prose into numbers the workspace can treat like any other time series. Analysts can build dashboards that start from structured management and macro signals rather than from a PDF and cross reference it with other data sources. On top of that, users can leverage AI agents to interrogate the underlying signals directly: asking why management tone deteriorated this quarter, which peers show the same pattern, or how a central bank's hawkish shift historically preceded moves in the curve. The agent doesn't have to read the transcript; it queries the indicator, pulls the context, and returns an answer grounded in the same numbers the dashboard is built on.
From calls and pressers to numbers you can trade on
Lexfi’s flagship products on OpenBB — Earnings Intelligence, Central-Bank Insights and Sentiment & Signals — sit on the same schema: standardised numeric indicators, point-in-time, refreshed daily. Every field is z-scored, so the same score means the same thing across companies, banks, and time. The signals are engineered to hold up under statistical scrutiny, not to sound good in a headline.
Earnings Intelligence covers >11,000 companies across 78 markets (96,612 scored calls), turning every quarterly call into 25 numeric indicators across five families:
- Quality (management, earnings, risk, growth, overall composite)
- Readability
- Sentiment
- Tone (uncertainty, confidence, litigious risk)
- Topic emphasis (capex/buybacks, hiring/ESG)
The evidence it carries signal: a long-only book holding the top quintile of the quality composite beats a country-matched benchmark by +10.4% a year, net of cost, at a t-statistic of 3.36 across 18,148 calls, and 59 of 100 formal tests survive false-discovery-rate control.
Central-Bank Insights covers 15 G15 central banks (627 press conferences since 2011), scoring each event into 11 indicators:
- hawkish-dovish index
- monetary-policy uncertainty
- inflation concerns
- forward-guidance clarity
- perceived growth
- inflation
- labor
- financial stability
- sentiment
- geopolitical risk
- labor-market tightness
Plus a categorical Hawkish/Neutral/Dovish label and two text fields (HIGHLIGHTS, CONCERNS). The hawkish-dovish index maps to roughly +1.1% of three-month forward equity return per standard-deviation move (~+94 bps per meeting), clearing multiple-testing control at q = 1.4e-5.
Sentiment & Signals fills the gap between scheduled events with the continuous flow of news, social, and market‑sentiment data. It includes:
- daily news‑tone, per ticker and FX, with ±1.5σ extremes pre‑flagged
- equity and crypto Fear & Greed with all seven components broken out
- five macro‑uncertainty indices: Economic Policy, Monetary Policy, Financial Market, Trade, and Geopolitical Risk.
- X bull/bear signals for a fast read on retail positioning
- Lexfi’s proprietary AI FX Ratio across USD/EUR/GBP/JPY/CHF/AUD, the same signal that powers one of their published strategies, with +5.46% alpha versus USD buy‑and‑hold and near‑zero equity beta.
The concrete payoff for a user
On any earnings day, you can rank every scored company by the quality composite in seconds, see which names moved the most on management or forward-looking sentiment relative to their own history, cross-reference against how the relevant central bank's hawkish-dovish score shifted the same week, and check whether news-tone z-scores flagged any ticker or FX pair that broke ±1.5σ that day. That's a joined view across corporate language, macro language and daily sentiment flow — at daily cadence, across 78 markets — that no single alternative-data feed structures at this scale.
Lexfi Quant Signals inside the OpenBB Workspace
Inside the OpenBB Workspace, the Lexfi Quant Signals app appears as three tabs – Earnings Calls, Central Banks, and Sentiment & Signals – each with its own set of synced widgets.
The Earnings Calls tab (5 widgets) covers 24 featured large caps: pick a ticker to see its call history, an AI-written insights panel (summary, AI scores, sentiment, digital strategy, ESG), sentiment and score trends across quarters, and the verbatim transcript. Everything is synced by ticker, and clicking a call drills straight to that event.
The Central Banks tab (4 widgets) opens on nine featured banks (Fed, ECB, BoE, BoJ, BoC, RBA, SNB, RBI, BCB) with an AI policy briefing, a color coded scorecard (stance, sentiment, inflation concern, labor tightness, policy uncertainty, guidance clarity), an economic assessment chart, and the full transcript with a link to the source video.
The Sentiment and Signals tab (11 widgets) opens with a KPI snapshot and a CNN Fear and Greed gauge. Then trends news tone for stocks and FX with z-score panels that flag statistical extremes, macro uncertainty indices, a net AI FX ratio per currency, X trading signals, and social and macro narrative pulses from X and Reddit. Every AI summary cites the exact transcript widget, so users can verify a claim against the raw text in one glance.
Because the AI agents connected to the OpenBB Workspace can read every table, chart, and dashboard as context, the app is AI-ready out of the box. You can ask "Summarize the latest NVDA call and flag the biggest risk," "Is the Fed more hawkish or dovish than last time?," or "Which tickers have the strongest bullish X signals right now?" and get a verifiable answer in seconds.
Lexfi API key
The app is free to use from the moment you connect it. An API key unlocks the full universe of data:
| Free (no key) | With X-Lexfi-Api-Key | |
|---|---|---|
| Tickers | 24 featured (dropdown) | Any global listed company ("Any ticker" field) |
| Central Banks | 9 featured (dropdown) | Any bank ("Any bank" field, e.g. BoK → 48 rows) |
| Sentiment History | 90 days (news/FX: YTD) | 365 days (news/FX: full trailing year) |
To add a key, set the custom header X-Lexfi-Api-Key in OpenBB's connection settings.
Use cases
FOMC workflow
On FOMC day, a macro and rates analyst opens Lexfi Quant Signals, sets the Central Banks tab to the Fed, and within minutes sees whether the hawkish drift is real and how the market is digesting it. The AI Policy Briefing gives the latest stance, the trend versus prior meetings, and the flagged concerns.
The Economic Assessment chart plots the Hawkish-Dovish Index across recent conferences; one dropdown overlays Inflation Concern to see if it is climbing, then Labor Market Tightness to check the jobs-market language. The Scorecard lines up every meeting side by side, and the Press Conference Transcript confirms the read in the Chair's own words, with a link to the source video.
A flip to the Sentiment and Signals tab shows the fallout: the AI FX Ratio for USD, monetary-policy and financial-market Uncertainty indices, and the News Sentiment and social pulse from X and Reddit.
Asked to tie it together, the OpenBB Copilot returns a sourced, cross-asset read in minutes: the 12-month tone has actually turned more dovish, but the latest meeting is still hawkish, with inflation concern and the removed forward guidance confirmed in the transcript, and markets digesting it as "hawkish but not disorderly," with a modestly USD-supportive tilt and no uncertainty shock.
Earnings‑day workflow
The morning after a print, an equity analyst covering semiconductors opens Lexfi Quant Signals, sets the ticker to AMD, and within minutes sees whether the beat is soft under the hood or just caught in a weak tape.
The AI Insights panel gives a plain-English summary of the call plus AI scores for growth, earnings quality, management tone, and risk. The summary flags cautious data-center guidance and a softer management-tone score. Clicking the call row loads the verbatim Transcript beside it, so the analyst reads the hedged guidance in management's own words. The AI Scores trend below shows the past trend across quarters, confirming management and risk scores have drifted lower for two calls rather than a one-off.
A flip to the Sentiment and Signals tab frames the reaction: the Ticker News Sentiment net tone, the Fear and Greed regime, and the X trading signals leaning short on semis.
With a Lexfi API key connected, the same routine runs on any company that the analyst is handed.
Get started
Open the OpenBB Marketplace and connect Lexfi Quant Signals to start for free.
To unlock any company, any central bank, and full-year history, get a Lexfi API key at app.lexfi.ai and connect it inside the Workspace.
