Skip to content
Convexity

Blog Company

2 min read

Introducing Convexity

One terminal for primary-source market data, an AI research layer, and the tools to act on both. Here is what we are building and why.

Convexity, a financial intelligence terminal

Convexity started with a simple frustration: serious retail research meant paying for five different tools and stitching their answers together by hand. A quote service here, a filings reader there, an options scanner, a macro dashboard, and a chat window open in a sixth tab. The data never agreed, and none of it could explain itself. We thought it should be one app, with one login, that shows the same number the same way everywhere.

What is inside

Convexity consolidates the stack a self-directed trader or independent analyst actually uses into a single terminal:

  • Primary-source market data: equities, options, and macro series, plus SEC filings, insider flow, 13F changes, and politician trades.
  • An AI research layer that reasons over that data and cites every claim back to its source.
  • Tools to act: a screener, an options lab, a convexity score, and a morning brief, all reading from the same pipe.
A heatmap of the data feeds Convexity consolidates
Forty-plus feeds across twenty-plus sources, behind one login.

The point of pulling everything behind one pipe is consistency. A figure on the dashboard, the same figure inside an AI answer, and the same figure on a detail page should all match, and each should tell you how fresh it is. That sounds obvious. It is also the thing that breaks first when data comes from five vendors with five refresh schedules.

The AI cites its work

The research layer is built around a rule we do not bend: every answer is grounded in primary sources, with the citation attached. You should never have to take a generated claim on faith. If the model cannot point to the filing, the print, or the series it used, it says so.

Freshness is a correctness property, not a nicety. When the assistant answers from data that synced an hour ago, it tells you the timestamp instead of implying the number is live.

The goal is not a chatbot bolted onto a data feed. It is a terminal that can explain itself, line by line, back to the source.

What is next

We are in pre-launch. Founder pricing is locked for life for early members, and the waitlist is open now. We will use this blog for product updates, the occasional market note, and the milestones worth marking. Thanks for reading, and for being early.

More dispatches

Every dispatch

  1. Market notes18 min read

    Missile contracts not yet final drove RTX's backlog jump

    RTX's defense backlog rose $45 billion in the third quarter to $164 billion, and about $37 billion of it was two undefinitized Tomahawk and AMRAAM contracts. The Pentagon obligated $154 million of AMRAAM's $20.7 billion ceiling at award.

  2. Market notes19 min read

    The ACA's 2027 price rise lands on full-price buyers

    ACA insurers asked for a median 15% premium rise for 2027. Tax credits absorb most of it for 87% of enrollees; buyers without one pay all of it. In 2026 there were 3.1 million of them, up from 1.9 million.

  3. Market notes21 min read

    Bending Spoons' real product is the next acquisition

    Bending Spoons grew revenue 126% last quarter and 3% organically. Inside the Milan firm Berkshire Hathaway just backed again: its playbook, cuts and debt.

See it in the terminal.

Primary-source data, an AI layer, and the tools to act on both. Founder pricing locked for life.