Downloads · 30 days
0
spreads-thesis
spreadsmoney/spreads-thesis
spreads-thesis is a machine learning model from spreadsmoney. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
Why an AI agent can earn the spread on tokenized stocks while the market sleeps.
Access
Public
Updated Sep 2, 2026
Repo size
—
Likes
0
Public
Repo makeup
Click a slice to open those files.
Other1.5 KB · 51%
At a glance
- License
- mit
- Access
- Public
- Created
- Aug 22, 2026
- Updated
- Sep 2, 2026
- SHA
- 17fc8ac3
Model card
From the Hugging Face model README
The Spreads Thesis
Why an AI agent can earn the spread on tokenized stocks while the market sleeps.
The issue. Robinhood's tokenized equities trade 24/7 on Robinhood Chain, but real price discovery happens on the primary markets between 9:30 and 4:00 Eastern, five days a week. Every night and every weekend, the passive liquidity left on chain is constant-product pools — formulas that cannot read news. They quote yesterday's price until someone trades them back into line. Prices go stale, honest makers leave, and the spread between what buyers pay and sellers get blows wide open.
The exploit. An agent estimates the true price from everything that still moves after hours — the close, futures, correlated names, headlines — and stands on both sides of it, slightly wider than its confidence. Every fill pays the half-spread. The one serious risk is adverse selection: quoting the old price to the first reader of a headline. Deciding "requote or stand down" is reading comprehension under time pressure — which is why this is an AI's trade, not a formula's.
Full documentation: https://spreads.money/docs/ · Kit: spreadsmoney/spreads-agent-kit · GitHub: https://github.com/spreadsmoney
Independent project; not affiliated with Robinhood Markets. Not financial advice.
- License
- mit
- Created
- Aug 22, 2026
- Updated
- Sep 2, 2026