About
Built by an ML engineer.Traded by Claude Code, with real money.
GammaRips ranks the US options market by liquidity every night and gives your AI a pool of roughly 40 to 50 liquid calls, plus the history to plan the exit. We use it ourselves. Claude Code reads the same pool through the same 9 MCP tools and trades it in a real-money agent account. Every trade is on the record, and traces of its sessions are public.
The live record
Claude Code trades the GammaRips pool with real money in an agent account. It sets a target and a stop before every entry. Every trade in the window counts, wins and losses.
trades closed at a profit
net realized
closed at the agent's pre-set target
median trade, on premium
N=14 real-money trades, entries Sep 14 to Oct 2, 2026. Realized broker fills.
Past results: real-money trades and historical pool data. Educational only. Not investment advice.
Who builds it
Evan Parra
Founder & builder
ML engineer and AI architect who ships production AI systems on Google Cloud. He built GammaRips end to end: the nightly scan, the liquidity funnel, the enrichment layer, the outcome history behind the exit lab, and the MCP server your AI connects to.
Also runs evanparra.ai for AI strategy and data integration consulting.
Claude Code
The live trader
Runs unattended on a VM, one bounded session per phase: preflight, entry, monitoring, and review. It reads the GammaRips pool through the same 9 MCP tools your AI gets, builds its own plan, sets a target and a stop before every entry, and trades a real-money agent account. Redacted traces of its sessions, with its reasoning quoted verbatim, are public in the open-source harness repo.
How the engine works
- 1After the close, the engine ranks about 3,500 optionable US stocks by liquidity.
- 2It keeps the 100 most liquid names, then the bullish ones.
- 3It selects one out-of-the-money call per name, on contract liquidity.
- 4Each trading morning it publishes the pool, roughly 40 to 50 contracts, with thesis, technicals, catalyst, and the outcome history behind the exit lab.
The result: median open interest of the pool contract is 4,705, against 906 before the liquidity rule (N=1,725 pool contracts Aug 24 to Oct 15, 2026, vs N=3,501 over the 60 scan days before the rule). Read the full method on the methodology page.
Why it's built this way
- ✓Liquidity first. The pool is selected on contract liquidity, so your AI starts from contracts you can enter and exit near the quote.
- ✓Your AI builds the plan. GammaRips serves data and tools, and your AI builds a plan for your account and your risk. Plans built from the data spread out instead of crowding into one shared contract.
- ✓The exit comes from history. The exit lab shows how often past pool contracts hit each profit level within 3 trading days, so the target and the stop are set on purpose.
- ✓Leakage-checked and documented. Every column carries its as-of boundary, and every filter and rule ships as a playbook your AI can read.
- ✓Receipts in public. The live record counts every trade in the window, wins and losses. The Lab publishes each experiment with its method, sample size, and verdict.
Trust & responsibility
GammaRips is a data vendor. It publishes options data, historical pool outcomes, and our own agent's real-money record, for education. Your AI builds the plan with you, and you decide and trade your own account. GammaRips never places trades for you and never manages your money.
Have a question? Reach out at evan@gammarips.com or use the form below.