Thursday links: MCPs and debt stories

Blockworks has an MCP

One way to understand the significance of the Model Context Protocol (MCP) is to think about it as a progression of who can do what on the internet.

First, the HTTP standard enabled humans to visually browse the web: Open Weather.com, for example, and see tomorrow’s forecast.

Later, APIs allowed software to access the web: The app on your phone can ask Weather.com for tomorrow’s forecast.

Now, the MCP standard allows LLMs to act on the web: “Hey agent: Check tomorrow’s forecast and if it’s sunny, send an email to my manager saying I’m not feeling well.”

You, hard-working reader, would never do that, of course. But you might want your agent to check some crypto prices?

If so, you could do that yourself by checking a website, thanks to HTTP. Or open an app on your phone that queries a database, thanks to APIs.

As of recently, though, you can also ask an LLM.

You probably won’t, though, because LLMs tend to hallucinate about things like that — unless they’re connected to a reliable source of data. This is what MCP is for.

First developed by Anthropic in November 2024, the Model Context Protocol is a set of rules (hence “protocol”) that standardize how LLMs connect to external tools and sources of data.

In the case of onchain data, you might ask Claude what the price of bitcoin has been doing lately. Claude can then query Blockworks’ MCP server (if you’re a subscriber), where it will see a menu of all the data Blockworks provides. Seeing that market data on bitcoin is of course available, Claude then requests recent prices. Blockworks delivers the data to Claude and Claude delivers it to you.

In whatever form you requested it.

The magic of connecting an LLM to a reliable source of data is that you can do pretty much anything with it. Price charts, sure. But much more analytical stuff, too, like comparing the fundamentals of one token to another. (Yes, some tokens really do have fundamentals.)

For example, I asked Claude to chart the net margins of the two largest onchain marketplaces for collectible trading cards:

I chose to share this one because it’s not something you’d expect to see in traditional investing: Courtyard makes twice as much off its users as Collector Crypto does — for the exact same service.

More importantly, connecting Claude to the Blockworks MCP makes me feel like I’m back in front of a Bloomberg terminal: all the data a trader could want and everything can be turned into a chart.

Except this is better because it’s much easier to use: Just ask for what you want, in plain English, and the LLM will build it for you. Even if you ask for an entire trading terminal — like the one my colleague built with Claude. In 30 minutes.

Connected to an MCP, your agent can just do stuff. Amazing.

More Wigglesworth

In my review yesterday, I focused on the colorful characters of Robin Wigglesworth’s new book, “A Fabulous Debt.” But that left out a lot of great anecdotes that deserve to be shared.

Like a bond offering made to London investors at the height of the 1820s investing bubble — a £1 million loan on behalf of the “Quicapouse” tribe of the Mississippi. The bankers said the money would be repaid over 700 years, which seems like a long time. Also, it would be paid “in shipments of wild geese,” Wigglesworth notes, “which would supposedly feather British beds and feed its masses.”

The Quicapouse (either fictional or an alternate spelling of Kickapoo, I’m not sure) do not appear to have found any buyers, unfortunately. But the fact that someone thought they might — and that The Courier newspaper thought it worth reporting — is indicative of the enthusiastic demand for bonds in the City of London at the time.

“Any kind of fantastical enterprise could try to raise money in London,” Wigglesworth adds, “such as an effort to drain the Red Sea to find the gold and jewels of the Egyptians drowned while chasing Moses.”

However excessive it sometimes got, London’s enthusiasm for investing was an enormous advantage for England. During the Napoleonic Wars, for example, Wigglesworth estimates that France would have had to pay a yield of 50% to borrow additional funds from the bond market. England, by contrast, could borrow at 6% or less.

It’s difficult to win a war against an enemy whose cost of capital is one-tenth of your own.

This helps explain why Napoleon sold Louisiana to the United States for the famously low price of $15 million: The price was effectively set at the maximum of what the US could borrow from London’s bond markets.

As cheap as the price seems now, the US government did not have $15 million to hand. So it paid by issuing bonds to France — which France accepted only because English investment banks had agreed in advance to purchase at a discount.

(Yes, England’s bankers were funding France’s war against England.)

The English learned the power of bond markets from the Dutch, who had an immense financial advantage over the Spanish during their long fight for independence. “The rates that the Spanish crown sometimes had to pay for its loans,” Wigglesworth writes, “were similar to those charged by Dutch pawn shops.”

The English noticed. Wigglesworth suggests the reason England deposed its home-grown King James II in favor of the Netherlands’ William of Orange was that England was borrowing at 8-14% at the time and the Netherlands was borrowing at 3-5%.

Such is the power of bond markets.

Wigglesworth similarly attributes the outcome of World War I to the US government’s ability to sell Liberty Bonds to Americans, the proceeds of which were used to keep England and France in the war while the US built a military. 

For most Americans, who bought strictly for patriotic reasons, Liberty Bonds were the first experience with fixed-income investing. They had a lot to learn. “At the beginning of the Liberty Bond drives,” Wigglesworth writes, “the Treasury would not infrequently receive letters from bond buyers asking when and where they would have to pay the interest on the bond.”

It’s still worth learning about, and “A Fabulous Debt” is a fun way to do it.

— Byron Gilliam

_{Brought to you by:}

Meridian 2026 is Stellar's annual gathering for the institutions, fintechs, and developers putting financial infrastructure onchain.

Join them October 28-29 at Convento do Beato in Lisbon for two days on tokenization, payments, and what it takes to run them in production.

Register today with Blockworks10 for 10% off.