◆ ALWAYS UP TO DATE · OUR MODEL KNOWS EVERY LATEST EARNINGS CALL

Ask anything about earnings calls.

Real questions, recorded 2026-08-18 and replayed word for word.

COMPANIES
12,000+
70 countries · all 11 GICS sectors
EARNINGS CALLS
252,000+
ask what any speaker said, back to 2020
MODEL TRAINING DATA
5+ years
discuss new calls within minutes

All sectors and industries covered

From oil majors to chip makers: if a company holds earnings calls, you can ask about it.

Information Technology
INFORMATION TECHNOLOGY
Health Care
HEALTH CARE
Financials
FINANCIALS
Industrials
INDUSTRIALS
Consumer Staples
CONSUMER STAPLES
Materials
MATERIALS
Utilities
UTILITIES
Real Estate
REAL ESTATE
Communication Services
COMMUNICATION SERVICES
ALL 11 GICS SECTORS · 74 INDUSTRIES

What you can ask it to do

The model interprets; every figure comes from a query, every quote is sourced.

FINANCIAL INTERPRETATION

Reads a quarter like an analyst: growth, margins and guidance, straight from the call.

SENTIMENT & TONE

Hears what numbers leave out: confidence, hedging and shifts in tone between quarters.

COMPARISONS

Two companies, one topic, side by side, each side backed by its own earnings calls.

TRENDS ACROSS QUARTERS

Follows one metric through call after call and returns the whole development, in percent.

SECTOR READS

An industry's entire quarter in one read: every call weighed, one verdict with quotes.

FOLLOW-UP DEPTH

A conversation with memory: ask, drill down, zoom out. It never loses the thread.

Testimonials

Our users put the chat to work on very different jobs, from a quick check to a deep dive.

“I ask for earnings call analysis across a whole sector's quarter and get one verdict with the quotes behind it. That used to be a week of transcripts.”
Markus T.
HEAD OF RESEARCH
“Twelve names on my watchlist, one earnings call summary each after the call. My earnings nights went from three hours to twenty minutes.”
Laura B.
PRIVATE INVESTOR
“I ask for a chart and it draws one from the quoted figures, axes and all. Nothing is computed, everything comes from the transcripts.”
Jonas K.
DATA ANALYST
“Revenue development across the last three quarterly earnings calls, in percent, in one answer, and each number carries its exact quote.”
Elena R.
EQUITY ANALYST
“Two companies, one topic, side by side, each side backed by its own calls. Basically my prep for positioning meetings.”
David S.
PORTFOLIO MANAGER
“Every claim names the speaker and the call it came from. It became my source for earnings calls; answers go straight into my notes.”
Werner H.
BUY-SIDE ANALYST
“It hears the tone shift between two quarters, the hedging in the guidance, the confidence. That's the part I used to listen for myself.”
Stefan W.
FUND MANAGER
“When the calls don't answer something, it says so instead of filling the gap. That honesty is why I quote it.”
Robert F.
FINANCIAL JOURNALIST
“I ask, then drill five questions deep, and it never loses the thread. Like chatting with an AI that read every earnings call transcript.”
Tim H.
RETAIL TRADER
“I ask for a table of what three companies said and paste it straight into my client notes. Formatted, sourced, done.”
Richard B.
WEALTH ADVISOR

Two plans, monthly, cancel whenever

Same AI model, same data foundation, only volume differs.

BASIC
24.99/ month
◆Around 750 questions a month
◆Ask, 3M tokens a month
◆The full data foundation
◆Archive search beyond it
◆Three languages
MORE ASK
PRO
39.99/ month
◆Around 3,000 questions a month
◆Ask, 12M tokens a month
◆Everything in Basic
◆Four times the budget
◆For daily use across many names

Payments processed by Paddle.com · Buyer Terms · Privacy · Refund

Frequently asked questions

Everything worth knowing before the first question.

A model trained to interpret earnings calls, with 252,000+ transcripts in its knowledge base and every new call added within minutes. You ask in plain language; the answer comes with verbatim quotes, queried figures and named sources. If you can ask a question, you can use it.

Never from the model, which interprets but does not invent. Every figure comes from a database query or a verbatim quote from a transcript, every answer lists the calls it draws on, and a claim that cannot be backed by a source does not make it into the answer.

New earnings calls land in the AI model's knowledge base within minutes of publication and are ready to discuss right away. The knowledge base reaches back to 2020, so a question can cover a single quarter or compare five years of calls side by side.

echat-1.0 is fast and made for summaries and quick lookups. echat-1.5 does deep reasoning on complex, multi-step questions and uses more tokens per answer. You switch between them under the input box, per question, and both work on the same knowledge base.

Each plan includes a monthly token budget. Every answer counts against it, deep reasoning counts more than a quick lookup, and the meter under the input box always shows where you stand. The budget resets with every billing period, unused tokens simply refresh.

Yes. Ask for a comparison or a trend and the answer attaches a bar or line chart or a table, drawn directly in the chat and built only from queried figures, never from generated numbers. Charts work on mobile too, with axes, tooltips and readable labels.

English, German and Spanish. Ask in your language and the answer follows it automatically, no setting required; quotes always stay verbatim in their original language, because a translated quote is no longer evidence. The interface comes in the same three languages.

Every answer has a copy button that exports it as markdown with the sources attached, ready to paste into a document, a chat or another model. Charts flatten to text bars and tables to markdown, so the shape of the data survives the trip into any other tool.

The answer says so. Every answer states how many passages it rests on and where they come from, and a gap is reported as a gap instead of being filled with a guess. That rule is deliberate: a missing number is annoying, an invented one is dangerous.

Yes. Both plans run monthly and can be cancelled whenever, with no notice period and no questions asked. They use the same model and the same data foundation; only the included volume differs, so switching up or down never changes the quality of an answer.