What does Cursorhood actually do?
Cursorhood acts as a local orchestration bridge between Cursor and Robinhood. It turns Cursor agents into autonomous trading bots running on Robinhood’s Agentic Trading MCP, using the markdown strategies and risk parameters you specify.
How do local risk hooks protect my account?
System prompts alone cannot guarantee an LLM won’t make an errant trade during market volatility. Cursorhood injects a local deterministic check in Python before any tool call reaches Robinhood. If an order exceeds your daily max loss, position size, contract limit, or minimum trade interval, the local hook denies the execution locally.
Where are my API keys and session tokens stored?
On your local machine. Cursorhood listens strictly on 127.0.0.1. Credentials and API keys are saved directly into your OS keyring (Windows Credential Manager on the Windows build, and the OS keyring on other systems). They are never written to unencrypted database files or sent to a Cursorhood server. Cycles still call Cursor with your key, and allowed orders still go to Robinhood.
Will running multiple bots burn through my Cursor API budget?
Large context windows and tight polling loops can consume Cursor usage quickly. Cursorhood mitigates this by isolating heavy context (news, filings, full price tape) from standard execution cycles. You can also assign lightweight models (like Composer or Auto) to rapid scanning bots while reserving larger models for specialized analysis.
How does paper trading work, and how realistic is it?
Simulation mode is turned on by default. Cursorhood includes a local paper broker that processes limit, stop, stop-limit, and market orders, accounting for current bid/ask quotes and maintaining local FIFO accounting for realized P&L. You can compare paper performance side-by-side with live snapshots before switching simulation off.
What are the limitations or edge cases with options trading?
Cursorhood currently supports single-leg options strategies. The local risk hook enforces contract count caps, max notional exposure, and days-to-expiry (DTE) windows. Local rules automatically reject assignment/exercise attempts and prompt the agent to close open contract positions instead.
How does Cursorhood handle Robinhood API rate limits?
Robinhood’s Agentic Trading MCP spans 79 tools. To prevent rate-limiting when running multiple bots simultaneously, each bot is constrained to its own asset class (equities, options, or crypto) and tool subset, with configurable sleep cycles outside US market hours.
Are Cursor and Robinhood accounts included in the purchase?
No. You must provide your own active Cursor subscription/API key and a funded Robinhood account with Agentic Trading access enabled. Cursorhood provides the software runtime to connect them. Sign up from the Required section.
What is the difference between the $25 Windows Beta and the MIT Version?
MIT Source (Free). The open-source Python repository is for developers who want to inspect, audit, or customize the code. It requires Python 3.11+ and manual virtual environment setup.
Windows Beta ($25 once). A standalone, pre-packaged executable bundling Python 3.11, the Cursor bridge, and launch scripts (Cursorhood.cmd) for quick, double-click deployment without terminal setup. The Windows terms cover this download.
Is Cursorhood offering financial advice or managing funds?
No. Cursorhood is strictly an open-source/standalone software utility running locally on your computer. It does not provide financial advice, hosted signal feeds, or custodial services. You maintain complete control and liability over all trading strategies and API credentials.