About aiSwingX™

We analyze US filings and write the news. Press releases and headlines carry the numbers a company wants shown. The notes and tables inside the 10-Q and 10-K hold different ones. We read the filing itself and work out what those numbers mean, how large they are against the company and the prior quarter, and why they matter — then write that analysis so an investor without an accounting background can follow it. Plain language is the delivery; analysis is the substance.

Who it's for

• Retail investors — you buy US stocks but could never get through a filing. One article takes three minutes.
• Brokerages, financial media, fintechs — filing coverage every earnings season that you can publish in your app or news platform as is. White-label and API/MCP ingestion supported.

Structure of an analysis article

Every article follows the same shape: headline and subhead, a two-to-three-sentence lead, then an At a glance box of four lines — one number, one analogy, one source, one thing to watch next quarter. The body runs "what happened → what the number means → why it matters → the company's view and context → what investors should watch".

For example, the SoFi Technologies (SOFI) article for Q2 2026 opens with "For every $100 it lent, SoFi's books say $104.60", then explains what that $4.60 is, why a fifth of the company's net worth is money it hasn't received yet, and which single number to check next quarter.

Each article closes with:

• A calculation table — every figure used in the article, with its value and where it sits in the filing (statement, note, MD&A), so readers can check the arithmetic themselves.
• A glossary and an FAQ — accounting terms explained with that article's own numbers, and the questions a reader would ask, answered from the filing.
• Related news links — recent press coverage of the same company: headline, outlet, date and link, so you can see how the market treated the filing without leaving the page. Headlines and links only; no article text is reproduced.
• A ticker page — the metrics in each article accumulate per ticker, quarter after quarter, so the same number can be followed over time.

Articles analyze and report what the filing says; they are unrelated to any decision to buy, sell or hold a security. Investment decisions are the reader's own; consult a qualified professional where appropriate.

Reliability

Article drafts are written by AI from the filing itself and reviewed before publication under a fixed protocol: the headline number must appear in the filing, the arithmetic must hold, the figures must match XBRL data, and the text must contain no trading recommendation or claim about intent. If any check fails, the article is not published. Every figure cites its source in the original filing; if an error is found after publication, we correct it and mark the correction.

Every article is published in Korean and English at the same time — not a translation, but the same filing written to the same structure, with identical figures and sources.

Sharing and view counts

Every page carries the same share bar at the top and bottom, with one-click sharing to X, Threads, Reddit, LinkedIn and Facebook (Naver on the Korean edition) and a copy-link button. At the right you will find the estimated reading time and the page's views: the figure beside the person icon counts visits by people, and the figure beside the robot icon counts reads by AI services such as ChatGPT, Claude and Perplexity. Search-engine bots and plain scripts are excluded from both. We show the two separately because articles are now read through AI assistants as well as by people.

Using the articles

Subscribers can copy an article or download it as .md/.txt. License customers ingest automatically through the API and MCP. Attribution is a condition of redistribution; white-label delivery without attribution is available by separate agreement. See plans →

For AI agents and developers

aiSwingX™ articles are not written for people alone. Every article is published simultaneously as a web page, a Markdown version (/p/<slug>.md) and structured data (NewsArticle, FAQPage, per-ticker metrics); /llms.txt and a news sitemap tell AI crawlers what is new. Ask a web-browsing AI assistant about a US filing and it can read and cite the same sourced facts a human reader sees. We count AI reads separately from human views and show both on every page.

• MCP server — connect aiSwingX™ to Claude, ChatGPT or any MCP-capable agent and ask "what did Broadcom disclose about lease guarantees?"; the agent returns the article, the four-line summary and the calculation table with sources.
• REST API — assemble articles from blocks (title, four lines, sections, images, share links, copyright); filter by ticker, sector, topic, importance and date. Paid sections return as locked blocks without a license.
• RSS (/feed.xml) and a briefs feed (/briefs.xml) — new articles and hourly market briefs from SEC and Federal Reserve sources.

Usage is documented in the API & MCP docs. When citing, include the article URL and the filing source.

Filing change score

Every article and filing carries a 0–100 score. It is not a credit rating, a price target or a buy/sell opinion. It measures, through nine published rules, how much the numbers reported in that filing moved against the comparison period: revenue change 15, net income change 15, earnings quality (operating cash flow ÷ net income) 15, inventory versus revenue 10, receivables ratio 10, debt and equity 10, auto flags 10, filing calendar 10 and editorial importance 15, capped at 100. Seventy and above reads "Large change", 40–69 "Moderate", below 40 "Small change".

Every input comes from the inline XBRL values in that same filing and the calendar items extracted from it. The article shows each component's points together with the figure behind it (for example, "operating cash flow / net income = 54%"), so you can check how the score was produced. A high score means the numbers moved a lot; it says nothing about the direction of the share price. Free readers see the score, the label and the names of the top three components; subscribers see all nine with their figures, and the API serves the same split.

Even where the score, its label and the calendar are visible without payment, they are for personal, non-commercial use. Showing them to your own customers or readers requires the License plan.

Improving

Where the change score asks how far the numbers moved, the Improving ranking asks which way. It compares figures a company printed in its own 10-Q or 10-K against the prior-year comparatives printed beside them in the same document. No outside estimates, no analyst consensus, no share prices, no volumes. All we do is hold up the same ruler the same way every quarter.

Ten lines are read: revenue 20, operating income 20, gross margin 15, cash from operations 15, free cash flow 10, net cash position 10, receivables ratio 5, inventory ratio 5. Lines a filing does not provide are dropped, and the result is normalised over the remaining points to a 0-100 scale. Every row carries both underlying figures — prior year and this filing — and a link to the EDGAR original, so you need not take our arithmetic on trust.

The trend beside a rank (improving, slowing, unchanged, weakening, first reading) is the direction of this score, not of a share price. It compares the same company's previous filing. As quarters accumulate, it answers a question of fact: how many quarters running have this company's numbers been getting better?

This ranking is not a price forecast and not investment advice. A company whose figures improved may see its shares fall, and the reverse is equally true. Appearing in the ranking is not a recommendation to buy, and absence from it is not a recommendation to sell. We do not know what share prices will do; we know what is printed in the filings. The full notice is in clause 6 of the disclaimer.

Free readers see the score, trend and reason for the top five places; subscribers see the whole ranking with the figures behind all ten components. The API (/api/v1/improving) and the weekly email follow the same split.

Filing calendar

Filings do not predict prices, but they are full of dated facts: debt maturities, warrant and lock-up expiries, when backlog will be recognized, buyback authorizations, litigation and regulatory dates, and events after quarter end. The Calendar page collects these sentences from the 10-Q and 10-K filings of covered tickers and orders them by date. Dates and periods are as the company wrote them, with nothing estimated; subscribers see each item's filing sentence and EDGAR link, and items stated as a period rather than a date are marked as such. New filings are added automatically, subscribers receive a weekly calendar email in their own language, and the data is available through the API.

Getting in touch

For partnerships, data licensing and redistribution, press citation, or the API, write to us through the contact page. No account is needed and our editorial team reads and replies. Email works too — aiswingx.com@gmail.com — in Korean or English. If you already subscribe, billing, account and article questions move faster through the support thread in your dashboard, where the history stays together.

Getting it by email

We write on the mornings a new filing analysis goes out, and send nothing on days there is none. It is not a daily newsletter; it is a letter we write when we have something to say.

A free registration brings the one number we went looking for per article, and a link. A subscription adds what to watch next quarter to the same email, drops the upsell, and sends a separate note the moment a filing scores 70 or more on our change score. Each kind can be switched on or off from your dashboard.

Metrics computed from the filing figures

Alongside the articles, some figures are computed by a single formula from the XBRL values companies file with EDGAR, and kept for every ticker we cover. They are material rather than interpretation: the formula and the source are all there is. They appear on ticker pages and through the API.

• Non-operating share of pre-tax income = (pre-tax income − operating income) ÷ pre-tax income — the part of a quarter's profit that came from outside operations, such as marks on stakes, interest and currency. Quarters with pre-tax income at or below zero, or under $10m, are left out because the ratio becomes noise; an operating loss is written as 100 with the reason stated.
• Principal due within twelve months = what the company itself tagged as long-term debt repayable in the next twelve months. Not our estimate.

Neither is a forecast, and a larger figure does not mean better or worse. They are recomputed daily.

We record the outcome, not just the date

A dated item is not finished when we write it down. Once the date passes we record what actually happened against the same item. Not every past date gets an outcome: a period marker such as "the next twelve months", or a dividend or interest payment date that executes itself, has nothing to resolve. We keep outcomes for the ones whose ending is news — trials, maturities, conversions, earn-outs. The record of what we said would happen, and what did, is the evidence we can offer.

What is free and what opens to subscribers

The free part carries what happened: the headline, the lead, the plain-language summary, one headline number and one comparison for scale, the first section and the FAQ. The subscriber part carries how large it is: the breakdown by year or line, ratios and multiples, the arithmetic table, the quarterly metrics and what to watch next. The filing calendar follows the same line — a free page shows the opening of each filing sentence, while the full sentence, the Korean translation, the EDGAR link and the complete list are for subscribers. A check enforces this boundary before anything is published.

On the site

• Search — articles, tickers and briefs in one query, by ticker, company, keyword or number.
• Save — the Save button on articles and cards keeps them in your dashboard to read again.
• Market briefs — SEC 8-K items and Federal Reserve releases collected hourly, newest first, on the landing page and the briefs page.
• New-article alerts — leave an email address, confirm via the link we send, and we notify you when a new article is published.

aiSwingX™ is not investment advice. It offers no recommendations or trading signals; investment decisions and their outcomes are the reader's own.