Introducing dibsresearch
Announcement · By dibsresearch ·
Ask about a company, a sector or the economy, and open the passage behind each number. Free while in early access.
Today we’re opening our Newsroom, and we’re starting with what dibsresearch is. You ask a question about a company, a sector or the economy in your own words, and it answers from SEC filings, company releases and official statistics, with the passage each figure comes from. While in early access, it’s free to use.
We built it because the facts an analyst needs are public, but scattered. A company’s guidance sits in its earnings release, the detail behind it in the 10-Q, the explanation in the call and the context in a statistic from the BLS or the Federal Reserve. Pulling those together by hand can take an afternoon, and an AI answer you can’t trace back to its source is hard to use.
So every figure in an answer has a footnote that opens the document at the passage it comes from. If an answer gives Netflix’s revenue forecast for 2026, $51.1 billion, its footnote opens Netflix’s letter to shareholders of 20 October 2026 at the sentence that states it. You can follow up as you would with a colleague, or open the Grid and ask the same question of a whole peer group at once.
Behind it is one index of more than 290,000 documents: SEC filings for more than 9,000 companies, the releases and presentations companies publish themselves, rules and court opinions, and official statistics from BLS, BEA, the Federal Reserve, the U.S. Treasury and Eurostat. Everything is read in full and searched together, so one search finds the 10-Q and the release that discuss the same number. Earnings calls we record and transcribe ourselves open in the company view, with the text running along with the audio.
On top of the documents we build data of our own. Company KPIs are read straight from the filings, each with the table cell it came from, and checked before they’re shown. When the Associate answers a question about a KPI, it looks in this data first and cites the cell.
The same data works outside dibsresearch, too. With dibsresearch MCP, ChatGPT, Claude, Cursor and other assistants that support MCP can look up KPIs, statements and filings, each with its source.
There are no prices, no broker research and no consensus estimates in dibsresearch. The numbers you see are the ones companies filed and agencies published, and each one opens the page it came from. Coverage follows what the SEC publishes, so it’s strongest for U.S. companies today.
We’re adding companies to our KPI data wave by wave, and older parts of the archive are still being filled in. There’s no daily limit and nothing to count while we build. Create an account and ask your first question.